Classification by Research Focus

Separating the reviewed papers based on the purpose of the research, we classified them into three categories: theory building, theory verifying, and applications. Theory building is second in order with a 14.1% contribution, theory testing is third with only 9.2%, and application papers are identified as the major purpose of research, accounting for 76.7% of all analyzed papers. Although there are many research papers focusing on theory building, the percentage is low compared to the overwhelming amount of application papers. Figure 2.11 shows an increase of papers on theory building until 2010, when the trend reached a peak and then proceeded to descend to 40%. However, theory-building papers did come back up in 2012 to a higher level than the one in 2010. Theory testing is consistently trending upwards while application papers are showing a positive trend, although the overall number of published articles in 2013 is going to still grow. Application papers focus on the practical implementation of technologies or ideas;

therefore we are seeing a large number of application papers as practitioners are analyzing SG for implementation. Authors seem to be focusing on the process of implementing those practical applications around SG, based on field and lab results without proper theory supported, but with a strong pragmatic support.

Based on the numbers shown on both Figures 2.10 and 2.11, we can reject hypothesis H2b, as we can see that the number of papers focusing on applications are now higher than the theory based ones. Our prognosis is that the applications’ papers will continue growing for a longer period of time, while the theory related ones will still climb until the reach the peak and begin descending.

Figure 2.10: Primary purpose of papers

Figure 2.11: Analyzed papers by research purpose 2.4.3.3 Classification by Data Collection Method

Empirical research papers can be classified by the data-collection method used. Almost one third of the papers used case studies (32.5%); Archival research accounts for 21.2% of all papers and laboratory research also accounts for 27.2%. It is not surprising to see case studies as the majority of data collection efforts because SGD technologies are recent and being fine-tuned in the field by simulating and quantifying their impact to a specific selected case. Later this case will be compared to others under different characteristics to validate the hypothesis and shape

theories. That is, case studies represent a grounded reality. Figure 2.13 shows that trends are positive in most cases.

Field and lab research are at the highest level in the period of time analyzed, showing increased interest in collecting data at these locations as real life validations are needed.

Considering all these facts, we support hypothesis H2c, as we see that case studies are the primary source of data collection, but as this data is available in time, the use of archival information will increase.

Figure 2.12: Research by collection approach

Figure 2.13: Papers by data collection categories & time

2.4.3.4 Classification by Data Analysis Technique

Papers can also be classified based on the numerical analysis techniques used. Although some of the published papers have no data analysis, we identified 12 techniques used in the analyzed papers. Some papers use more than one technique, but for analytical purposes we chose the one technique which is the focus of the paper. Figure 2.14 weights the relative importance of the techniques used when measured as a percentage of the total incidences (966).

Figure 2.14: Data analysis techniques used

Simulation is the analysis technique used most frequently in the analyzed papers with a 35% contribution and continuing on an upward trend. We expected to find such a large number of simulation papers because 45.8% of the case studies studying applications use simulation as the main tool and 17% of the simulations use the Monte Carlo method. Papers without data (22%) are also showing an upward trend as new techniques, opportunities and challenges are constantly discovered and discussed before conducting any data analysis.

Figure 2.15: Data analysis techniques evolution in time

2.4.3.5 Classification by Discipline

To separate the categories of papers, we used the six layers presented by Giannfranco Chicco (2010) identified on the basis of interaction between elements in the Smart Grid papers. It is important to emphasize that there are going to be overlaps as two or more layers can be

addressed in a single paper.

Table 2.3: Disciplines and Descriptions (Chicco, 2010) Category Description of Category

Physical (P) Network interconnections used for the distribution of energy

Regulatory (R) Applying standards, rules and other types of incentives or penalties to moderate the activity in the market, control prices and/or protect consumers Environmental

(E)

Activities to have an impact on the defense of human life and protection of the planet

Social (S) Customers’ willingness to do something by investing on devices, changing consumption habits, participating in the utility companies’ programs, etc.

Economics (F) Those activities that have an impact on the economy of the consumer or the company

Information (I) Includes efforts to integrate ICT in the distribution of energy via management, communications, and control

The Smart Grid literature has been evolving from an early focus on transmission and generation with unilateral flow of communications and electricity to bidirectional communications whereby electricity flows from the generator to the consumption location and from the consumption location distributed generator to the grid. Using the definitions in Table 2.3, we identified the papers with the six categories using the first letter as, identifier, with the exception of economics changed to financial (F) to avoid repeating (E) which was already being used for environmental papers. All categories show upward trends to date.

Figure 2.16: Distribution of papers by category

Figure 2.17: Number of papers by type in time

An important finding on this chart is the increase of economics related papers in the past three years. These papers tended to focus on efficiency because the energy generation and conversion process are of low efficiency (Dovì et al., 2009). This financial analysis has not peaked yet, and is rapidly growing to overtake second place from physical interconnection papers. The papers show an increased awareness on the financial impact of SG with the appearance of a new business model based on the elements of distributed generation, electric vehicles and automated metering that are inspiring new customer behaviors. The role of the consumers and producers merging together as prosumers will bring added complexity to the commercial relationship between utilities and entities generating energy for self-consumption because they can also sell their excess capacity to the utility company. In the same way that communication and energy now are flowing bi-directionally instead of unilaterally, energy costs will also have tariffs from and for the utility company, creating a new situation that has not been addressed recently. This new model is an effect of the upward trend of economics papers related to SG because it will require participation of many stakeholders who were previously passively involved in this process. The social and cultural changes from this new business model have to be analyzed in detail before companies fail to involve consumers and implement SG with their strong opposition. On the other hand, if companies successfully involve the stakeholders in the enhancement of the generation, distribution and consumption processes, the result could bring forth a new quality of life for consumers in general. Chapter 4 presents this enhanced model that has been developed within this dissertation.

2.4.3.6 Taxonomy of Paper’s Purpose

By analyzing the general purpose of the published paper, we use several categories to

Supply Chain Management, Environmental Issues, Service Design, Process Design, Scheduling, Planning, and Blackouts. Strategic papers will present clearly defined strategies or guidelines for future implementation. Quality Management focuses on optimization and performance. Supply Chain Management centers on the continuous supply of inputs. Environmental Issues papers are concerned with reducing environmental impacts (“greener” planet). Service and Process Designs are very similar because energy provision is a service, but the focus on the generation, transmission and distribution process will be considered as process design, while service design is related to the activities around the process to aid and enhance the customer perception;

Scheduling is related to Planning in time; while the Planning category is used for those activities as developed beforehand. Finally, Blackouts papers concentrate on avoiding power outages. The results are presented in Figure 2.18 and their trends are presented in Figure 2.19. The three most important topics in the reviewed literature are Strategy (31%), quality management (29%), and supply chain (14%). It is not surprising to see strategy as the leading category because the alternatives for proper implementation are still being developed in some of the technologies.

Figure 2.18: Categories of papers by topics

The value in quality management is increasing due to the growing importance of using

typical loss of generated energy due to energy transportation losses (Görbe et al., 2012), which in the US account for 5-12% (Lauby, 2010) and 15 to 20% in India (Sinha et al. 2011). The third important category is the supply chain, which describes how energy users have changed from only being consumers to now becoming “prosumers” who generate and consume energy (Zhabelova & Vyatkin 2012).

Figure 2.19: Categories by time

Figure 2.19 shows how Quality Management is overtaking strategy in 2013 in total number of papers, which confirms the increasing notion of a SG definition as a way to look for efficiency (Wissner, 2011). Strategy is still strong because some areas of SG have not been fully developed, such as electric vehicles. EVs bring new and expensive interactions including driving range, recharging time (Silvester et al. 2012) and vehicle charging, which still needs to be considered before massive implementation can be accomplished (Shuaib, 2012). Based on the numbers and trends, we reject hypothesis H2d because we see that strategies have been passed by quality.

2.4.3.7 Smart Grid Technologies

There are several technologies under the umbrella of Smart Grid Distribution (SGD); the most typical ones are Advanced Metering Infrastructure (AMI), Demand Response (DR), Distribution Automation (DA), Electric Vehicles (EV), Distributed Generation (DG), System Efficiency Improvement, Self-Healing, Cybersecurity, and Distributed Storage. Looking at the 966 papers, we classified them into 11 categories. Figures 2.20 and 2.21 show distribution and trends. Only efficiency and Volt/VAR analysis show reduction of papers beyond the 2010 level, all others show positive trends as techniques are becoming more popular and important to our society.

The three most important technologies based on the number of papers are: Distributed Generation (23%), Distribution Automation (18%), and System Efficiency (16%). These three areas show the current trends of SGD. Distributed Generation is a growing area mostly in Europe, where there is a major effort to move to self-generation using renewable resources that could reduce 2-4% of the energy losses in Europe (Lasseter, 2011); based on this growth we support hypothesis H2e.

There is also an important effort to automate the distribution process with the use of ICT devices. The automation process will allow quicker and better responses on energy consumption that will result in efficient use of energy, thus reducing the negative impact on environment (Koroneos at al. 2012). Additional methods to make energy distribution more efficient vary from replacing transformers with new more efficient ones (Grigoras et al. 2010) to the identification of the most probable configuration of the wind to locate mills (Chertkov et al. 2011).

Figure 2.21: Trend of SGD technologies in time

2.4.3.8 Originating Countries

Figure 2.22 shows the countries where the papers originated, i.e. the home of the first authors, the country referred in the paper, or the sponsor country of the publication. More than one quarter of the papers are from the US, which generated 27% of all literature, followed by China with 13%, Canada with 6%, Italy with 5% and India with 4%. Figure 2.23 shows the continuous increase of papers coming from other countries not listed in the top 5. The diverse composition of this category shows a worldwide interest in the area. The US is still maintaining

hypothesis H2f, which states that the US is the world leader in peer-reviewed SGD literature generation. Something important to emphasize is the growth of literature from other countries.

Figure 2.22: Distribution of papers by country

Even though the top 5 represent 55% of all papers analyzed, the number of contributions from other countries is increasing, as there is growing worldwide attention on this subject. Some countries are specializing in certain technologies. For instance, Portugal is writing more papers about electric vehicles than most countries, with a particular focus on particle swarm for vehicle to grid scheduling (e.g. Soares et al. 2012, Sousa et al. 2012).

It is interesting to note that as many as 81 countries have participated in the generation of literature. We see a technological wave throughout the world to meet the upcoming challenges and to be prepared for implementation. Environmental protection and energy sustainability are two of the many subjects addressed globally and their importance is critical.

Analyzing matrices by doing cross-tabulation between columns and rows, we have identified some facts that are important for this area of knowledge. Table 2.4 shows that applications have the major contribution for the published papers with a 77% contribution, followed by theory building in the modeling and analytics area. We are seeing more application papers of Smart Grid Distribution because this area is being optimized with the aid of technologies already available and previously used for transmission (e.g., SCADA, D-SCADA) or in the areas of communication (e.g. wireless technologies). The empirical analysis is the largest share of the pie with 60%, followed by the conceptual papers with less than half of the empirical contribution. 92% of the conceptual papers are focused on practical applications.

Table 2.4: Row & Column Percentages for Research Classification versus Focus

Table 2.5 shows that case studies is the data collection method used by the largest group of papers with 33% of all analyzed articles. Archival research is used mostly for applications (89%) because secondary data sources are used to represent the way things are in real life, not to

Application Theory

Empirical Research 58% 68% 72% 60% 73% 16% 11% 100%

Conceptual & General 30% 10% 6% 25% 92% 6% 2% 100%

Modeling & Analytical 11% 21% 22% 14% 63% 22% 15% 100%

Survey 1% 0% 0% 1% 100% 0% 0% 100%

SG Education 0% 0% 0% 0% 100% 0% 0% 100%

Total 100% 100% 100% 100% 77% 14% 9% 100%

papers using case studies, followed by laboratory research with around 29 and 37% of the papers. These numbers show that using implementations in the field and case studies represent 51% of all papers confirming the expected focus on practical applications for future implementation validation.

Table 2.5: Row & Column Percentages for Data Collection versus Paper’s Focus

Table 2.6: Row & Column Percentages for Data Analysis vs. Paper’s Focus

Table 2.6 shows that simulation is the technique used most often with 36% of the total Application Theory

Archival Research 25% 11% 9% 21% 89% 7% 4% 100%

Field Research 17% 24% 21% 18% 71% 19% 11% 100%

Surveys 1% 0% 0% 1% 100% 0% 0% 100%

Simulation 34% 40% 45% 36% 73% 16% 12% 100%

No data analysis 25% 15% 11% 22% 86% 9% 5% 100%

Time Series 21% 19% 17% 21% 80% 13% 8% 100%

Descriptive 13% 13% 15% 13% 75% 15% 10% 100%

Model 4% 3% 6% 4% 75% 11% 14% 100%

the field need prior validation to prevent catastrophic failures. The creation of algorithms to simulate consumer patterns was also used to build theory, while some other papers simulated current or expected conditions to test already developed theory. Simulations represent a larger share than theory testing (45%), theory building (40%) and applications (34%).

Table 2.7: Row & Column Percentages for Conference papers’ SG technologies vs. Purpose

Tables 2.7 and 2.8 show the relationship between the Smart Grid technology and the focus of the conference and journal papers. Table 2.7 shows the relationship for conference publications only. As shown in Table 2.7, strategy is the most published paper focus followed closely by quality management and supply chain. Strategies for distribution automation (25%) and efficiency (19%) represent 44% of all published strategies. This is due to the new definition of the Smart Grid as the automation of distribution with the inclusion of ICT for optimized results. The focus of conference papers for distributed generation is most importantly related to supply chain (70%), as this concept of home generation of own energy is growing rapidly in

StrategyQuality

Dsgn Planning Others Total StrategyQuality Mgt

automation (26%), but the approach is beyond strategy and moving into the “how to” mode. The process design papers focus mainly in distribution automation (31%), demand response (15%) and efficiency (9%) because these are the practical papers— “the nuts and bolts” of the Smart Grid implementation.

Table 2.8: Row & Column Percentage for Journal Papers’ SG technologies vs. Purpose

Table 2.8 shows the relationship between the Smart Grid technology and the focus of journal papers. As shown in Table 2.8, the order of the conference papers’ contribution is changed for journals, because quality is the dominant focus here, followed closely by strategy and supply chain. Quality related papers focus mostly on efficiency (22%) and distributed generation (20%). SGD Strategies at journals have the same influence of distribution automation and efficiency (21%) and they are 42% of all published strategies. The focus of journal papers for distributed generation is most importantly related to supply chain (68%). Planning papers are focused on achieving efficiency (33%) and distribution automation (22%). Service design papers

StrategyQuality

Dsgn Planning Others Total StrategyQuality Mgt

advanced metering (10%). Journal papers show technologies reaching maturity because conference papers emphasize strategy while journals are paying more attention to Quality.

Table 2.9: Row & Column Percentages for Country vs. Paper’s Purpose

Table 2.9 shows the relationship between the focus of the papers and the country where the document was generated. As the US is the leader in strategy, quality, and supply chain, we support hypothesis H2f, stating that the US is the leader in strategic papers. As shown in Table 2.9, strategy is the most published paper focus with 28% followed closely by supply chain (28%), and then quality (23%). Although the US is the leader in all categories, it is important to emphasize the contribution of China in planning (16%), service (17%), and quality (14%), while Canada has a balanced contribution on most topics.

Table 2.10 shows the relationship between the SG technologies and the countries publishing articles. The leadership of the US in the Cybersecurity (44%), Demand Response (37%), and the top 3 (25%) shows the interest on developing these technologies while the rest of the world is following the example. Germany is writing more papers on efficiency (33%) than the rest of the countries, which shows an interest in migrating to efficient ways of generating

StrategyQuality

Dsgn Planning Others Total StrategyQuality Mgt

contributor for electric vehicles, with 9% of all papers on this subject. In general, we see that the US is leading over all SGD technologies with China as a distant second. Based on this analysis we can conclude that the US is and will be the leader of all clean energy efforts that come to place in the next 20 years.

Table 2.10: Row & Column Percentages for Country vs. SG Technology

2.4.4 Further Analysis and Prognosis

DG DA Eff. DN EV CS AMI DR Others Total DG DA Eff. DN EV CS AMI DR Others Total USA 25% 25% 25% 15% 20% 44% 24% 37% 36% 27% 21% 17% 15% 4% 5% 10% 5% 8% 14% 100%

China 11% 14% 16% 13% 19% 15% 11% 9% 7% 13% 20% 20% 20% 8% 11% 7% 5% 4% 6% 100%

Canada 7% 4% 5% 8% 6% 5% 5% 6% 6% 6% 28% 13% 13% 11% 7% 6% 6% 6% 11% 100%

Italy 7% 6% 3% 8% 4% 2% 2% 4% 6% 5% 30% 22% 10% 12% 6% 2% 2% 4% 12% 100%

India 3% 6% 5% 5% 3% 3% 5% 4% 4% 4% 17% 24% 17% 10% 5% 5% 7% 5% 10% 100%

Korea 5% 3% 5% 1% 3% 8% 0% 2% 3% 4% 29% 17% 20% 3% 6% 14% 0% 3% 9% 100%

UK 1% 5% 2% 5% 9% 3% 4% 4% 4% 4% 9% 24% 9% 12% 18% 6% 6% 6% 12% 100%

Germany 1% 1% 7% 0% 6% 6% 5% 4% 4% 3% 9% 6% 33% 0% 12% 12% 9% 6% 12% 100%

Brazil 3% 5% 1% 9% 3% 0% 5% 0% 2% 3% 23% 26% 6% 23% 6% 0% 10% 0% 6% 100%

Spain 4% 2% 2% 4% 0% 2% 9% 0% 0% 2% 33% 17% 13% 13% 0% 4% 21% 0% 0% 100%

Japan 4% 2% 3% 0% 3% 0% 0% 4% 3% 2% 38% 13% 21% 0% 8% 0% 0% 8% 13% 100%

Iran 2% 2% 1% 0% 3% 2% 0% 6% 5% 2% 20% 20% 5% 0% 10% 5% 0% 15% 25% 100%

Others 27% 24% 26% 31% 23% 11% 29% 22% 22% 25% 25% 18% 17% 10% 7% 3% 7% 5% 10% 100%

Total 100% 100% 100% 100% 100% 100% 100% 100% 100% 100% 23% 18% 16% 8% 7% 6% 6% 6% 11% 100%

To understand the impact of neighboring countries as blocks, we are consolidating the number of publications into continents, in order to assess the influence and spread of these technologies across borders. We are using the country of the first author only, classified into the main 5 continents: Africa, America, Asia, Europe and Oceania.

Figure 2.25: Journal papers by continent and year

Both Figures 2.24 and 2.25 show the analyzed papers grouped by publishing outlet and the continent of the paper’s first writer. America is the leader in conference papers that present strategies and technologies earlier, but Europe has been very close to taking leadership in conference papers’ production. With these facts we cannot support hypothesis H2g, which states that America is the leader in conference papers if we compare continents instead of countries. It

Both Figures 2.24 and 2.25 show the analyzed papers grouped by publishing outlet and the continent of the paper’s first writer. America is the leader in conference papers that present strategies and technologies earlier, but Europe has been very close to taking leadership in conference papers’ production. With these facts we cannot support hypothesis H2g, which states that America is the leader in conference papers if we compare continents instead of countries. It