Uncertainty in the learning rates of energy technologies: An experiment in a global multi-regional energy system model

The diffusion of promising energy technologies in the market depends on their future energy production–cost development. When analyzing these technologies in an integrated assessment model using endogenous technological learning, the uncertainty in the assumed learning rates (LRs) plays a crucial role in the production–cost development and model outcomes. This study examines the uncertainty in LRs of some energy technologies under endogenous global learning implementation and presents a floor-cost modeling procedure to systematically regulate the uncertainty in LRs of energy technologies. The article narrates the difficulties of data assimilation, as compatible with mixed integer programming segmentations, and comprehensively presents the causes of uncertainty in LRs. This work is executed using a multi-regional and long-horizon energy system model based on “TIMES” framework. All regions receive an economic advantage to learn in a common domain, and resource-ample regions obtain a marginal advantage for better exploitation of the learning technologies, due to a lower supply-side fuel-cost development. The lowest learning investment associated with the maximum LR mobilizes more deployment of the learning technologies. The uncertainty in LRs has an impact on the diffusion of energy technologies tested, and therefore this study scrutinizes the role of policy support for some of the technologies investigated.     

Global technology learning and national policy—An incentive scheme for governments to assume the high cost of early deployment exemplified by Norway

In this paper it is argued that technology learning may be both a barrier and an incentive for technology change in the national energy system. The possibility to realize an ambitious global emission reduction scenario is enhanced by coordinated action between countries in national policy implementation. An indicator for coordinated action is suggested. Targeted measures to increase deployment of nascent energy technologies and increasing energy efficiency in a small open economy like Norway are examined. The measures are evaluated against a set of baselines with different levels of spillover of technology learning from the global market. It is found that implementation of technology subsidies increase the national contribution to early deployment independent of the level of spillover. In a special case with no spillover for offshore floating wind power and endogenous technology learning substantial subsidy or a learning rate of 20% is required. Combining the high learning rate and a national subsidy increases the contribution to early deployment. Enhanced building code on the other hand may reduce Norway’s contribution to early deployment, and thus the realization of a global emission reduction scenario, unless sufficient electricity export capacity is assured.     

Endogenous implementation of technology gap in energy optimization models—a systematic analysis within TIMES G5 model

Evaluation of global diffusion potential of learning technologies and their timely specific cost development across regions is always a challenging issue for the future technology policy preparation. Further the process of evaluation gains interest especially by endogenous treatment of energy technologies under uncertainty in learning rates with technology gap across the regions in global regional cluster learning approach. This work devised, implemented, and examined new methodologies on technology gaps (a practical problem), using two broad concepts of knowledge deficit and time lag approaches in global learning, applying the floor cost approach methodology. The study was executed in a multi-regional, technology-rich and long horizon bottom-up linear energy system model on The Integrated MARKAL EFOM System (TIMES) framework. Global learning selects highest learning technologies in maximum uncertainty of learning rate scenario, whereas any form of technology gap retards the global learning process and discourages the technologies deployment. Time lag notions of technology gaps prefer heavy utilization of learning technologies in developed economies for early reduction of specific cost. Technology gaps of any kind should be reduced among economies through the promotion and enactment of various policies by governments, in order to utilize the technological resources by mass deployment to combat ongoing climate change.     

Implications of technological learning on the prospects for renewable energy technologies in Europe

The objective of this article is to examine the consequences of technological developments on the market diffusion of different renewable electricity technologies in the EU-25 until 2020, using a market simulation model (ADMIRE REBUS). It is assumed that from 2012 a harmonized trading system will be implemented, and a target of 24% renewable electricity (RES-E) in 2020 is set and met. By comparing optimistic and pessimistic endogenous technological learning scenarios, it is found that diffusion of onshore wind energy is relatively robust, regardless of technological development, but diffusion rates of offshore wind energy and biomass gasification greatly depend on their technological development. Competition between these two options and (existing) biomass combustion options largely determines the overall costs of electricity from renewables and the choice of technologies for the individual member countries. In the optimistic scenario, in 2020 the market price for RES-E is 1 €ct/kWh lower than in the pessimistic scenario (about 7 vs. 8 €ct/kWh). As a result, total RES-E production costs are 19% lower, and total governmental expenditures for RES-market stimulation are 30% lower in the optimistic scenario.     

Technology learning for renewable energy: Implications for South Africa's long-term mitigation scenarios

Technology learning can make a significant difference to renewable energy as a mitigation option in South Africa's electricity sector. This article considers scenarios implemented in a Markal energy model used for mitigation analysis. It outlines the empirical evidence that unit costs of renewable energy technologies decline, considers the theoretical background and how this can be implemented in modeling. Two scenarios are modelled, assuming 27% and 50% of renewable electricity by 2050, respectively. The results show a dramatic shift in the mitigation costs. In the less ambitious scenario, instead of imposing a cost of Rand 52/t CO₂-eq (at 10% discount rate), reduced costs due to technology learning turn renewables into negative cost option. Our results show that technology learning flips the costs, saving R143. At higher penetration rate, the incremental costs added beyond the base case decline from R92 per ton to R3. Including assumptions about technology learning turns renewable from a higher-cost mitigation option to one close to zero. We conclude that a future world in which global investment in renewables drives down unit costs makes it a much more cost-effective and sustainable mitigation option in South Africa.     

Renewable energy technologies for irrigation water pumping in India: projected levels of dissemination, energy delivery and investment requirements using available diffusion models

Using the past diffusion trends of four renewable energy technologies for irrigation water pumping in India (SPV pumps, windmill pumps and biogas/producer gas driven dual fuel engine pumps), results of an attempt to project their future dissemination levels, have been presented in this study. The likely contribution of the renewable energy options considered in the study to the projected energy demand for irrigation water pumping in India has been estimated. Estimates of the associated investment requirements taking into account the learning effect have also been presented.     

Useful models for simulating policies to induce technological change

Conventional top-down and bottom-up energy–economy models have limitations that affect their usefulness to policy-makers. Efforts to develop hybrid models, that incorporate valuable aspects of these two frameworks, may be more useful by representing technologies in the energy–economy explicitly while also representing more realistically the way in which businesses and consumers choose between those technologies. This representation allows for the realistic simulation of a wide range of technology-specific regulations and fiscal incentives alongside economy-wide fiscal incentives and disincentives. These policies can be assessed based on the costs required to reach a goal in the medium term, as well as on the degree to which they induce technological change that affects costs over long time periods.     

A new generation of AI: A review and perspective on machine learning technologies applied to smart energy and electric power systems

The new generation of artificial intelligence (AI), called AI 2.0, has recently become a research focus. Data‐driven AI 2.0 will accelerate the development of smart energy and electric power system (Smart EEPS). In AI 2.0, machine learning (ML) forms a typical representative algorithm category used to achieve predictions and judgments by analyzing and learning from massive amounts of historical and synthetic data to help people make optimal decisions. ML has preliminarily been applied to the Smart Grid (SG) and Energy Internet (EI) fields, which are important Smart EEPS representatives. AI 2.0, especially ML, is undergoing a critical period of rapid development worldwide and will play an essential role in Smart EEPS. In this context, this study, combined with the emerging SG and EI technologies, takes the typical representative of AI 2.0—ML—as the research objective and reviews its research status in the operation, optimization, control, dispatching, and management of SG and EI. The paper focuses on introducing and summarizing the mainstream uses of seven representative ML methods, including reinforcement learning, deep learning, transfer learning, parallel learning, hybrid learning, adversarial learning, and ensemble learning, in the SG and EI fields. In this survey, we begin with an introduction to these seven types of ML methods and then systematically review their applications in Smart EEPS. Finally, we discuss ML development under the big data thinking and offer a prospect for the future development of AI 2.0 and ML in Smart EEPS. We conduct this survey intended to arouse the interest and excitement of experts and scholars in the EEPS industry and to look ahead to efforts that jointly promote the rapid development of AI 2.0 in the Smart EEPS field.     

Modeling technological learning and its application for clean coal technologies in Japan

Estimating technological progress of emerging technologies such as renewables and clean coal technologies becomes important for designing low carbon energy systems in future and drawing effective energy policies. Learning curve is an analytical approach for describing the decline rate of cost and production caused by technological progress as well as learning. In the study, a bottom-up energy-economic model including an endogenous technological learning function has been designed. The model deals with technological learning in energy conversion technologies and its spillover effect. It is applied as a feasibility study of clean coal technologies such as IGCC (Integrated Coal Gasification Combined Cycle) and IGFC (Integrated Coal Gasification Fuel Cell System) in Japan. As the results of analysis, it is found that technological progress by learning has a positive impact on the penetration of clean coal technologies in the electricity market, and the learning model has a potential for assessing upcoming technologies in future.     

The resilience of the Indian economy to rising oil prices as a validation test for a global energy–environment–economy CGE model

This paper proposes to test the global hybrid computable general equilibrium model Imaclim-R against macroeconomic data. To do so, it compares the modeled and observed responses of the Indian economy to the rise of oil price during the 2003–2006 period. The objective is twofold: first, to disentangle the various mechanisms and policies at play in India's economy response to rising oil prices and, second, to validate our model as a tool capable of reproducing short-run statistical data. With default parameterization, the model predicts a significant decrease in the Indian growth rate that is not observed. However, this discrepancy is corrected if three additional mechanisms identified by the International Monetary Fund are introduced, namely the rise in exports of refined oil products, the imbalance of the trade balance allowed by large capital inflows, and the incomplete pass-through of the oil price increase to Indian customers. This work is a first step toward model validation, and provides interesting insights on the modeling methodology relevant to represent an economy's response to a shock, as well as on how short-term mechanisms – and policy action – can smooth the negative impacts of energy price shocks or climate policies.     

Limits to leapfrogging in energy technologies? Evidence from the Chinese automobile industry

Limits to leapfrogging in energy technologies? One of the most attractive notions in the field of sustainable energy development is the concept of energy-technology “leapfrogging”. Leapfrogging through international technology transfer can be especially problematic because often developing countries do not have the technological capabilities to produce or integrate the advanced energy technologies themselves. Until they have acquired the capabilities to produce the advanced technologies themselves, most late-industrializing countries buy their new technologies from industrialized countries, usually through licensing or joint-venture arrangements. Empirical case studies of the three main Sino-US passenger-car joint ventures reveal that until the late 1990s, little energy or environmental leapfrogging occurred in the Chinese automobile industry as the result of the introduction of US automotive technology. An improvement in Chinese capabilities and more stringent Chinese energy and environmental policies are needed to induce energy leapfrogging in the Chinese automobile industry. Foreign firms also have a social responsibility to contribute to China's sustainable industrial development. In order to realize the promise of the leapfrogging, the limits to leapfrogging must be identified and acknowledged so that strategies can be devised to surmount the barriers to the introduction of advanced energy technologies in developing countries.     

Technology learning in a small open economy—The systems,modelling and exploiting the learning effect

This paper reviews the characteristics of technology learning and discusses its application in energy system modelling in a global–local perspective. Its influence on the national energy system, exemplified by Norway, is investigated using a global and national Markal model. The dynamic nature of the learning system boundary and coupling between the national energy system and the global development and manufacturing system is elaborated. Some criteria important for modelling of spillover ¹ are suggested. Particularly, to ensure balance in global energy demand and supply and accurately reflect alternative global pathways spillover for all technologies as well as energy carrier cost/prices should be estimated under the same global scenario. The technology composition, CO₂ emissions and system cost in Norway up to 2050 exhibit sensitivity to spillover. Moreover, spillover may reduce both CO₂ emissions and total system cost. National energy system analysis of low carbon society should therefore consider technology development paths in global policy scenarios. Without the spillover from international deployment a domestic technology relies only on endogenous national learning. However, with high but realistic learning rates offshore floating wind may become cost-efficient even if initially deployed only in Norwegian niche markets.     

Incorporating carbon capture and storage technologies in integrated assessment models

Low-carbon emitting technologies are a key component of technical change in integrated assessment models. We develop a methodology for incorporating technologies into computable general equilibrium economic models and demonstrate this methodology by implementing carbon capture and storage technologies in the MIT Emissions Prediction and Policy Analysis (EPPA) model. Three primary implementation issues are discussed: characterization of the technical system, translation of bottom–up engineering information into an economic model, and the depiction of realistic technology adoption rates. The specification of input substitution, relative costs, and plant dispatch are the most critical factors in technology representation. Technology adoption rates in economic models are governed by exogenous and endogenous constraints. A comparison of the current approaches used in economic models with the theoretical and empirical factors affecting adoption rates highlights opportunities for refining the current methods.     

Public support for energy sources and related technologies: The impact of simple information provision

Increasing public awareness and understanding of alternative energy sources and related technologies is an essential component of informed decision-making regarding new options of generating energy for a low carbon future. The current study examined the influence of psychological factors (i.e., pro-environmental beliefs, and subjective norms) and the provision of factual information on public support for a range of energy sources and related technologies. A representative sample of 1907 Australians completed an on-line survey that measured perceptions of a range of climate change and energy issues. Results showed that support for renewables is stronger than support for traditional fossil-fuel based energy sources (i.e., coal or gas) or nuclear energy. The provision of factual information about generation cost and emissions significantly changed support ratings, particularly when cost information was provided. Regression analyses revealed that pro-environmental beliefs were significantly related to support ratings for alternative energy sources. Subjective norms, however, were the strongest positive explanatory factor, suggesting that social mechanisms may be key drivers of support for new and emerging energy sources and related technologies.     

Discount rate for technology assessment: An application to the energy sector

When comparing the relative merits of technologies in the energy sector, analysis of a long time span is frequently required. Discounting of costs and benefits is usual in such analysis, and the choice of an appropriate discount rate can be vital. The available empirical evidence suggests that the appropriate real discount rate could vary widely depending on which concept it is based. For Australia, a value reflecting the social time preference rate could be as low as 2%, while one based on the social opportunity cost of capital would be considerably higher, probably 7–10%. A multiperiod linear programming model, MARKAL, suitable for energy technology assessment, was used to analyse the effect on the optimal long-term energy strategy for Australia, of using discount rates over the range 2–10%. The results indicate that capital-intensive technologies such as coal liquefaction and solar water heaters are disadvantaged when using high discount rates. This could result in energy strategies being pursued in which the level of oil imports is much     

Near-optimal energy management for plug-in hybrid fuel cell and battery propulsion using deep reinforcement learning

Plug-in hybrid fuel cell and battery propulsion systems appear promising for decarbonising transportation applications such as road vehicles and coastal ships. However, it is challenging to develop optimal or near-optimal energy management for these systems without exact knowledge of future load profiles. Although efforts have been made to develop strategies in a stochastic environment with discrete state space using Q-learning and Double Q-learning, such tabular reinforcement learning agents’ effectiveness is limited due to the state space resolution. This article aims to develop an improved energy management system using deep reinforcement learning to achieve enhanced cost-saving by extending discrete state parameters to be continuous. The improved energy management system is based upon the Double Deep Q-Network. Real-world collected stochastic load profiles are applied to train the Double Deep Q-Network for a coastal ferry. The results suggest that the Double Deep Q-Network acquired energy management strategy has achieved a further 5.5% cost reduction with a 93.8% decrease in training time, compared to that produced by the Double Q-learning agent in discrete state space without function approximations. In addition, this article also proposes an adaptive deep reinforcement learning energy management scheme for practical hybrid-electric propulsion systems operating in changing environments.     

Innovation and technology transfer through global value chains: Evidence from China's PV industry

China's success as a rapid innovation follower in the infant Photovoltaic (PV) industry surprised many observers. This paper explores how China inserted itself into global clean energy innovation systems by examining the case of the solar PV industry. The paper decomposes the global PV industrial value chain, and determines the main factors shaping PV technology transfer and diffusion. Chinese firms first entered PV module manufacturing through technology acquisition, and then gradually built their global competitiveness by utilizing a vertical integration strategy within segments of the industry as well as the broader PV value chain. The main drivers for PV technology transfer from the global innovation system to China are global market formation policy, international mobilization of talent, the flexibility of manufacturing in China, and belated policy incentives from China's government. The development trajectory of the PV industry in China indicates that innovation in cleaner energy technologies can occur through both global and national innovation processes, and knowledge exchange along the global PV value chain.     

Development of machine learning models to enhance element-doped g-C3N4 photocatalyst for hydrogen production through splitting water

Elemental doping has been widely adopted to enhance the photoactivity of graphitic carbon nitride (g-C₃N₄). Correlating photocatalytic performance with experimental conditions could improve upon the current trial-and-error paradigm, but it remains a formidable challenge. In this study, we have developed machine learning (ML) models to link experimental parameters with hydrogen (H₂) production rate over element-doped graphitic carbon nitride (D-g-C₃N₄). Material synthesis parameters, material properties, and H₂ production conditions are fed to the ML models, and the H₂ production rate is derived as the output. The trained ML models are effective in predicting the H₂ production rate using experimental data, as demonstrated by a satisfactory correlation coefficient for the test data. Sensitivity analysis is performed on input features to elucidate the ambiguous relationship between H₂ production rate and experimental conditions. The ML model can not only identify important features that are well-recognized and widely investigated in the literature, which supports the efficacy of the developed models but also reveals insights on less explored parameters that might also demonstrate significant impacts on photocatalytic performance. The method described in the present study provides valuable insights for the design of elemental doping strategies for g-C₃N₄ to improve the H₂ production rate without conducting time-consuming and expensive experiments. Our models may be used to revolutionize future catalyst design.     

Business models for model businesses: Lessons from renewable energy entrepreneurs in developing countries

Against the background of mounting research suggesting entrepreneurship as a means of increasing the uptake of renewable energy technologies (RETs) in developing countries, this paper presents the findings of an exploratory investigation into the business models used by renewable energy entrepreneurs in such countries. Forty-three entrepreneurs were interviewed in 28 developing countries and secondary information about country and regional conditions was analysed. We chose the Business Model Canvas as an analytical tool and the findings shed new light on established renewable energy business types. Three different types of businesses were identified – Consultants, Distributors, and Integrators; yet, there is also some overlap between these types. These business types appeared to parallel the life cycle progression of the business, but this requires further research. A key component of the study was to assess whether the types of businesses were related to country-level conditions to assess the impact of regional differences. These comparisons revealed consistencies between country-level characteristics and the entrepreneurs’ choice of business model. Conclusions suggest that different regions may support certain business models more than others due to differing levels of government interest in renewables, governance and policy support and the relative ease of doing business.     

Machine learning predictive models for optimal design of building-integrated photovoltaic-thermal collectors

This research article aims to examine the feasibility of several machine learning techniques to forecast the exergetic performance of a building-integrated photovoltaic-thermal (BIPVT) collector. In this regard, it uses multiple linear regression, multilayer perceptron, radial basis function regressor, sequential minimal optimization improved support vector machine, lazy.IBK, random forest (RF), and random tree approaches. Moreover, it implements the performance evaluation criteria (PEC) to evaluate the system's performance from the perspective of exergy. The use of these approaches serves the identification process to realize the relationship between the input–output parameters of the BIPVT system. The novelty of this work is that it utilizes and compares multiple learning algorithms to predict the PEC of BIPVT through design parameters. Hence, the research considers the parameter (PEC) as the essential output of the BIPVT collector, while the input parameters are the length, width, and depth of the duct, located under the PV modules, as well as the air mass flow rate. The results of the research for the statistical indexes of mean absolute error, root mean square error, relative absolute error (%), and root relative squared error (%) show values of (0.2967, 0.3885, 1.8754, and 1.5237) and (0.4957, 0.8153, 2.9586, and 2.8289), respectively, for the training and testing datasets. While R² ranges (0.9997-0.9999) for those datasets. Therefore, to estimate the exergy performance of the BIPVT collector, the RF model is superior to other proposed models.