Determinants of Residential Consumers’ Acceptance of a Utility-Scale Battery Energy Storage System in Malaysia: Technology Acceptance Model Theory from a Different Perspective

Abstract

In a developing country such as Malaysia, studies of determinants which influence residential consumers of the Battery Energy Storage System (BESS) are limited. This paucity of studies was the catalyst for this study and its aim to investigate the factors affecting acceptance by Malaysian residential consumers of BESS as it relates to the Technology Acceptance Model Theory. A sample of 331 residential consumers indicated that consumer attitudes, social norms and self-efficacy, or the perception of behavioral control, had a positive and significant relationship with the intention to use BESS. Additionally, trust was a factor that had a significant effect on the consumers’ perceptions of cost, benefits and anticipated effects. All these variables significantly affect consumer attitudes. These findings provide important insights into BESS and facilitate the development of policies and practices relating to BESS in developing countries such as Malaysia.

1. Introduction

Battery Energy Storage Systems (BESS) have garnered considerable attention in developing countries such as Malaysia due to their potential for energy savings, improved reliability, grid resilience, generation resource integration and assistance in reducing environmental impacts. Recently, Malaysia announced the introduction of a utility scale for BESS in order to achieve a total capacity of 500 MW between 2030 and 2034 in association with the country’s energy transition plan 2021 to 2024. This will be obtained by having 40% of renewable energy generation being provided by Solar PV generation [1]. Presently, there are more than 10 million customers in the Malaysian electricity market, of whom more than 8 million are residential consumers [2]. Taking this into account, the government of Malaysia has implemented numerous initiatives to increase Solar PV generation. These initiatives have included the Feed in Tariff (FiT), Large Scale Solar (LSS) and Net Energy Metering (NEM) [3].

Along with the rapid expansion of solar PV generation capacity, technology such as the Battery Energy Storage System (BESS) is growing in importance in its ability to shift power generation from the middle of the day, when there is frequently a peak in demand, to the evening [4,5]. Lithium-ion batteries have become more effective and cheaper in recent years. They are now viable options for storing and shifting at least a few hours’ worth of solar energy when required. As there is great potential for solar PV in the Malaysian market, there is a need to investigate consumer acceptance of BESS so as to increase the reliability and effectiveness of solar PV technology in Malaysia.

1.1. Motivations

Previous studies have focused on the acceptance of Malaysian household consumers of solar PV (e.g., [6]), but there is little in the way of published research that examines the role of the residential consumer in the growing household acceptance of BESS. To help address this shortfall, this paper reviewed the determinants of consumers’ acceptance and the use of the BESS in Malaysia and examined the role of the consumer in the emerging utility-scale BESS market. This paper employed a quantitative research survey to identify the determinants of consumers’ acceptance and the use of BESS by asking key questions of relevant communities. The findings enhance understanding of the drivers or determinants of consumer acceptance of BESS in Malaysia and provide important insights for policy makers and practices. This survey also provides a key to better understanding of consumer motivations as the energy market transitions, and the survey helps inform policy and strategic decision making aimed at achieving optimal integration of the technology.

1.2. Related Works

Recently, the variety of energy sources and the role that BESS plays has become of great interest to researchers because BESS is considered to be one of the technologies that helps reduce greenhouse gasses emissions [7]. Additionally, BESS facilitates the production, consumption and storage of energy from the main and subsidiary networks for consumers at the community and household level. It also contributes to lessening the congestion experienced by electrical networks and provides the general power system with various services in terms of both transmission and distribution networks [8]. In light of the benefits, responsible bodies from developed countries have installed BESS in their local energy projects [9]. These projects cater for various stakeholders such as householders, communities and various commercial sectors [10].

Green technology, such as BESS, is an important facility which can store extra or excess energy from the electrical generation system. This, in turn, optimizes the use of fossil fuels in energy supply which, ultimately, can reduce greenhouse gas (GHGs) emission produced by electrical generation. BESS enables a reduction in peak demand by leveling out demand during the peak hours, thereby increasing the efficiency and load factor, and, importantly, reducing the need to build new energy power plants to support the extra energy demand that occurs during peak usage times. BESS comprises user-friendly equipment that can be installed close to neighborhood areas, it provides cost savings for domestic electrical users by avoiding maximum charges incurred during peak hours as well as providing a solution to operational risks such as power outages [11]. More broadly, BESS can contribute significantly to sustainable energy development by enabling energy utilities to meet the growing energy needs associated with the country’s long-term economic growth, but, importantly, in an affordable, reliable and secure manner, while simultaneously providing a pathway to net zero carbon emissions.

Large-scale energy storage technologies have been developed with the intention of facilitating peak load reduction and load shifting [12]. Energy storage using the lithium-ion battery is one the major electrochemical energy storages [13]. This type of BESS has very high round-trip efficiency (around 95%, a low self-discharge rate and a high energy density). Medium-sized BESS can be installed to store electric energy via a home solar system using panels and a home battery; this enables consumers to entirely disconnect from the network [14]. Interestingly, studies such as [15] have provided a technical specification of the capability of a solar PV panel to supply the needed energy for at least one total daily recharge for a fully electric vehicle. Figure 1 illustrates the schematic diagram of a battery energy storage system and Table 1 sets out the benefits of BESS as presented in previous literature.

Figure 1. Schematic Diagram of a Battery Energy Storage System. Source: http://www.amdcenergy.com/battery-energy-storage-system.html (accessed on 15 May 2022).

Table 1. Example of Benefits of BESS.

No	Benefits	Author
1.	BESS is one of the important technologies in energy management as it facilitates consumers’ reduction of energy losses and improves energy efficiency in various energy systems.	[16]
2.	BESS can be one of the solutions to the problem of high and irregular voltage problems which arise due to the access of renewable energy sources by distribution networks.	[17]
3.	BESS can be used to maintain energy systems’ stability, reduce power quality problems in micro-grid systems and match demand with supply.	[18]
4.	BESS has environmental benefits by reducing the use of fossil fuel sources to meet the energy demand, thus reducing the release of carbon dioxide emission into the atmosphere.	[19]
5.	BESS has become an important technology for renewable energy resources and, with its availability, can increase the promotion of renewable energy resources such as solar energy and others.	[15]

Although it is very efficient, the main problems faced by this type of BESS relate to its short lifespan and high establishment costs [13]. Many manufacturers have increased the volume of BESS production in order to reduce the material cost which, ultimately, can reduce BESS’s price [20]. For instance, electrical vehicle manufacturers produced a half million cars by 2020 which led to a reduction in battery costs of 30%. Additionally, subsidies for BESS, implemented by several governments, have resulted in an increase in uptake [16]. For small-scale energy requirements, batteries are now the most popular and efficient power storage method. The electric grid must become more spatially and temporally flexible, and battery energy storage can be a crucial component [21].

1.3. Contributions

The purpose of this paper is to investigate the part played by trust, cost, risk, benefit, anticipated affect, attitude, social norms and perceived behavioral control in relation to the adoption of BESS by Malaysian households by analyzing the respondent data through Smart PLS 3.0. The findings of the study resulted from large-scale questionnaire surveys covering every region of Malaysia. These questionnaire surveys provide useful insights for the government and responsible bodies when formulating policies that aim to increase BESS usage. In terms of a comprehensive contribution to knowledge, the intention of this study is to fill the following research gaps. Firstly, this paper, by applying a comprehensive framework of the Technology Acceptance Model (TAM), examined how trust, positive affects (cost, risk, benefit and anticipated affect), outcome evaluations (attitudes and social norms) and perceived behavioral control affect consumer intention to use BESS. Presently, these strands of research do not exist in the literature for developing countries such as Malaysia. Secondly, integrating the trust variable in this framework provides insights into how trust plays a role in shaping the consumer’s intention to use BESS. Thirdly, this study contributes to policy and strategic decision making for countries undergoing energy market transitions by offering empirical information on the factors that influence consumer acceptance of BESS.

1.4. Organization of the Paper

The paper is arranged in the following way. In Section 1 the introduction of study including motivations, related works and contributions is presented. Section 2 contains the literature review and hypothesis development including a review of consumer acceptance of BESS, theoretical background and hypothesis development. Section 3 discusses the research method and the research context. Section 4 provides the analysis of results and the hypotheses testing. Section 5 discusses the results, and lastly, Section 6 covers the conclusions and future works.

2. Literature Review and Hypotheses Development

2.1. Review of Consumer Acceptance of BESS

The studies of consumer acceptance of BESS have focused primarily on developed countries such as Germany [22,23], Canada [24], the United Kingdom [25,26] and the Netherlands [21]. Data related to this acceptance are presented in Table 2. According to Table 2, the majority of the studies found that communities accepted the BESS. Nevertheless, this acceptance is dependent on several factors such as trust on the part of relevant stakeholders, the cost and technical safety. However, environmental values have not influenced consumer acceptance of BESS.

Table 2. Previous studies of Consumer Acceptance of BESS.

No	Authors	Sample	Methods/Framework	Variables	Findings
1	[22]	1247 public in Germany, distributed through social networks	Technology Acceptance Framework [27,28] Questionnaire Survey and Structural Equation Modelling	IVs: Perceived problems (5 items); Trust in industry (3 items); Trust in municipalities (4 items); and environmental identity (3 items) Mediating Variables: Perceived effects (7 items); and affect (1 item) DVs: General acceptance (1 item) and Local acceptance (1 item)	Significant: Trust in industry Trust in municipalities Perceived problems (only for general acceptance) Insignificant: Environmental identity
2	[23]	About 607 consumers in Germany	Online survey (474), focus groups (45), questionnaire (neighborhood—53), workshop (neighborhood—35)	Motives for storage acquisition; Perceived risks (not tabulated); Ownership preference; Willingness to participate; Residential vs. Community installations; Conditions preference for community ES; Trust in stakeholders; and potential energy provision and storage services	The interest in energy storage services is strongly dependent on the consumers’ perceptions of the costs versus potential savings.
3	[25]	46 British public	Deliberative workshops	Awareness and salience; Perceived risks and benefits—Aesthetics and spatial impacts; efficiency, environment and sustainability; reliability; safety; technological progress; Fairness; Independence; and Convenience.	Storage technologies are viewed as ambiguous, and acceptance is likely to be conditional on whether they can be designed, regulated, and governed in ways that address technical concerns about safety, environmental impacts and reliability while also addressing societal desires for equity and the protection of vulnerable groups.
4	[24]	1022 Canadian adult population	Online Survey	DV: Intention to support IVs: Positive effect; Perceived benefits; Social norms; Attitude; Costs (justified); Self-claimed knowledge; Problem perception; Perceived risks; Trust in developers; Environmental values; Negative affect; Climate change; Self-efficacy; and Costs (better spent elsewhere)	Significant and positive: Positive effect Perceived benefits Social norms Attitude Costs (justified) Self-claimed knowledge Problem perception Significant but negative: Perceived risks
5	[26]	1044 UK adult population	Same as above	Same as above	Significant and positive Positive effect Attitude Benefit Costs (justified) Trust in developers Problem perception (marginal) Significant but negative Self-claimed awareness Environmental values
6	[29]	Employees from the network operator’s numerous departments and roles, the Tegenstroom’s director, and directors of renewable energy cooperatives in the Netherlands (17 semi-structured interviews)	Semi Structured Interview	Societal benefits Societal costs Barriers and opportunities—policies and political power; organizational logic and sector structure; knowledge base; user practices; cultural significance; technology and infrastructure.	The stakeholders involved have different perceptions of the benefits of the neighborhood battery concept, and their expectations are not always aligned.

2.2. Technology Acceptance Framework (TAF) of Sustainable Energy Acceptance

This study used the Technology Acceptance Framework (TAF) of sustainable energy acceptance developed by [27]. The comprehensive framework is based on three theories including the Technology Acceptance Model [30], the Theory of Planned Behavior (TPB) [31] and Theories of Affects [32]. TAM explains user motivation factors such as perceived ease of use and usefulness and attitude towards the behavioral intention of technology usage [30,33]. TPB explains how an individual’s behavioral intention is influenced by their attitude, subjective norms and perceived behavioral control. Theories of Affect explain how individual attitude is influenced by the anticipated effect of certain forms of technology [32]. Based on a comprehensive framework, this study examines the relationship between:

• The consumer’s trust in BESS and their combined effects of affect and cognition (the perceived cost, perceived risk, perceived benefit and anticipated affect) of BESS;
• The consumer’s combined effects of affect and cognition (the perceived cost, perceived risk, perceived benefit and anticipated effect) of BESS and their attitude towards BESS;
• The consumer’s outcome evaluation (attitude and subjective norms) and perceived behavioral control towards BESS and their intention to use BESS in the future. The research framework of this study, adopted from a comprehensive framework of sustainable energy acceptance developed by [27], is shown in Figure 2.
  

  Figure 2. Research Framework.

2.3. Hypothesis Development

2.3.1. Consumer’s Trust in BESS and the Combined Effects of Affect and Cognition (the Perceived Cost, Perceived Risk, Perceived Benefit and Anticipated Affect) on BESS

Consumers’ trust in BESS’s developer and responsible bodies have a strong influence on the individual’s acceptance of BESS in a situation in which the individual is not familiar with new technology [34]. If an individual has a high level of trust in the energy regulator or developer, they will have a high level of acceptance of perceived cost, risk and benefits and their affective responses towards new technology such as BESS [35]. In the context of the BESS study, [22] it was found that trust in municipalities and industry had a positive impact on an individual’s perceived effects and the anticipated affect of public acceptance of stationary battery systems in Germany. On the other hand, a study by [24] found that trust in the developer was not important in terms of intention to use energy storage systems in Canada.

Hence, these beliefs form the basis for the following hypotheses:

H1a. 

Consumer’s trust in BESS is positively related to their perceived cost of BESS.

H1b. 

Consumer’s trust in BESS is positively related to their perceived risks of BESS.

H1c. 

Consumer’s trust in BESS is positively related to their perceived benefits of BESS.

H1d. 

Consumer’s trust in BESS is positively related to the anticipated effects of BESS.

2.3.2. Consumers’ Combined Effects of Affect and Cognition (Perceived Cost, Perceived Risk, Perceived Benefit and Anticipated Effect) of BESS and Their Attitude towards BESS

Perceived effect can be regarded as an individual’s belief related to technology in terms of its cost, risk and benefits [28]. The beliefs and anticipated effect related to certain technology will influence people’s attitudes towards the technology. In terms of the context of a study of BESS, [24] the study found that the perceived benefit and cost of BESS, as well as anticipated effects of BESS, had a positive significant relationship with individual acceptance of energy storage systems in Canada; a study by [26] in the UK came to the same conclusion. Research by [22] found that perceived effect has a positive significant relationship with general acceptance of stationary battery energy storage, while affect has a significant influence on local acceptance of stationary battery energy storage. Hence, the following hypotheses were proposed:

H2a. 

Consumer’s perceived cost of BESS is positively related to their attitude towards BESS.

H2b. 

Consumer’s perceived risk of BESS is negatively related to their attitude towards BESS.

H2c. 

Consumer’s perceived benefit of BESS is positively related to their attitude towards BESS.

H2d. 

Consumer’s anticipated affects of BESS are positively related to their attitude towards BESS.

2.3.3. Consumer’s Outcome Evaluation (Attitude and Subjective Norms) and Perceived Behavioral Control towards BESS and Their Intention to Use BESS in the Future

From the perspective of the Technology Acceptance Framework (TAF) developed by [27], three determining factors that influence individuals to act on certain green energy technologies such as BESS included: (1) attitude towards BESS based on an evaluation of the behavior considered either positive or negative, (2) the subjective norm, i.e., how an individual thinks about certain behavior based on the consideration of surrounding people, and 3) perceived behavioral control, i.e., how easy or difficult individuals believe it is to make the behavior happen. A study by [24] in Canada and [26] in the United Kingdom found that consumers’ perceived benefits, social norms and attitudes have a significant influence on their acceptance of BESS. Hence, the following hypotheses were proposed:

H3. 

The consumer’s attitude towards BESS is positively related to their acceptance of BESS.

H4. 

The consumer’s social norms associated with BESS are positively related to their acceptance of BESS.

H5. 

The consumer’s self-efficacy or perceived behavioral control towards BESS is positively related to their acceptance of BESS.

3. Methods

This study utilized a purposive sampling design. The target population for this research was Malaysian residential electricity consumers from five regions in Malaysia, including (1) Central Region (Selangor/Putrajaya/Kuala Lumpur); (2) Northern Region (Perlis/Kedah/Pulau Pinang/Perak); (3) Southern Region (Negeri Sembilan/Melaka/Johor); (4) East Coast Region (Kelantan/Terengganu/Pahang) and; (5) Borneo Region (Sabah/Sarawak/Labuan). The cross-sectional online survey was conducted in the period between June and September 2021. A link to the survey that was deployed in Google Forms was delivered to Malaysian residential electricity consumers via a Facebook community group page. The anonymous survey questionnaire took up to 10–15 min to complete, once participants agreed to take part. The G*Power statistical software was utilized to determine that for a desired power of 0.80, effect size f² of 0.15 (medium), and eight predicting variables, the sample size necessary for this study is 109 [36]. The 331 respondents that participated in this study were deemed to have enough statistical power to analyze the relationship between the variables of study [37]. Smart-PLS v3 software was used to assess the relationship between the selected constructs by Partial Least Squares-Structural Equation Modeling (PLS-SEM), as used in prior studies [38]. PLS-SEM is an important analytical technique for testing the theoretical framework from the prediction perspectives [39].

To operationalize the variables, the study adapted scales already validated and had shown good explanatory power in the previous literature. The adapted questionnaire consisted of three parts. The first part entailed the understanding of battery energy storage systems. The second part is comprised of a demographic profile. The third part was devoted to constructing the main variables and thus detailing the items related to the independent and dependent variables. The questionnaire was anchored on a five-point Likert scale where “strongly disagree” was represented by 1, and “strongly agree” was denote by 5. The measurement items of the variables us in this study was adapted from [7,24,26].

4. Results

4.1. Sample Characteristics

Table 3 sets out the demographic profile of participants in which 53.8% were female. The highest proportion of participants were middle-aged (33.5%), aged between 31 and 40 years, followed by the 20–30 year old cohort (31.7%) and those aged between 41 and 50 years at 26.3%. Additionally, 261 (78.9%) of participants were Malay/Bumiputera/Sabah/Sarawak, 49 (14.8%) participants were Chinese and 19 (5.7%) were Indian. Of these participants, 48.6% had bachelor degrees and 25.7% of the participants had a PhD or equivalent. Furthermore, 36.6% held professional positions. The majority of the participants were from the education and training industry which constituted a total of 137 participants (41.4%). A total of 44.4% of the participants had monthly household incomes between ($1154.58 to $2608.59 (M40). Most of the participants lived in Selangor/Putrajaya/Kuala Lumpur (64.7%), followed by the second major residency areas of Perlis/Kedah/Pulau Pinang/Perak at 20.8%. The most current type of residence was a terrace house (48.0%), followed by an apartment type at 12.4%. An urban area (62.2%) constituted the largest part of the sample of their current settlement area, followed by a suburban area (19.6%) and a rural area (18.1%). The household size was 4 to 6 persons (61.3%) or 203 participants, followed by 0 to 3 (23.3%) and 14.5% of households had 7 to 9 occupants.

Table 3. Profiles of the respondents.

Gender	Frequency	Percentage	Industry	Frequency	Percentage
Male	153	46.2	Agriculture, Food and Natural Resources	8	2.4
Female	178	53.8	Architecture and Construction	16	4.8
Age	Frequency	Percentage	Arts, Audio/Video Technology and Communication	5	1.5
Less than 20	4	1.2	Education and Training	137	41.4
20–30	105	31.7	Government and Public Administration	25	7.6
31–40	111	33.5	Hospitality and Tourism	3	0.9
41–50	87	26.3	Information Technology	17	5.1
51 and above	24	7.3	Manufacturing	44	13.3
Ethnicity	Frequency	Percentage	Other	76	23
Malay/Bumiputera/Sabah/Sarawak	261	78.9	Estimated Monthly Income (Household)	Frequency	Percentage
Chinese	49	14.8	B40 (Earning below $1154.34)	114	34.4
Indian	19	5.7	M40 ($1154.58 to $2608.59)	147	44.4
Other	2	0.6	T20 (Earning more than $2608.59)	70	21.1
Education Status	Frequency	Percentage	1 USD = 4.2015 MYR (Average currency in 2020)
Secondary School	8	2.4	Residency	Frequency	Percentage
Certificate	5	1.5	Selangor/Putrajaya/Kuala Lumpur	214	64.7
Diploma/Equivalent	21	6.3	Perlis/Kedah/Pulau Pinang/Perak	69	20.8
Bachelor’s Degree/Equivalent	161	48.6	Kelantan/Terengganu/Pahang	30	9.1
Master’s Degree/Equivalent	49	14.8	Negeri Sembilan/Melaka/Johor	17	5.1
PhD/Doctoral Degree/Equivalent	85	25.7	Sabah/Sarawak/Labuan	1	0.3
Other	2	0.6	Current Settlement Area	Frequency	Percentage
Current Position	Frequency	Percentage	Rural Area	60	18.1
Professional	121	36.6	Suburban Area	65	19.6
Top Management	11	3.3	Urban Area	206	62.2
Middle Management	37	11.2	Current Type of Residence	Frequency	Percentage
Supervisory	12	3.6	Village House/Long House	10	3
Administrative or Clerical	20	6	Flat	18	5.4
Technical	23	6.9	Apartment	41	12.4
Housewife	5	1.5	Condominium	34	10.3
Retiree	4	1.2	Townhouse	6	1.8
Full Time Student	85	25.7	Terraces	159	48
Other	13	3.9	Semi Detached	19	5.7
Household Size	Frequency	Percentage	Cluster House	6	1.8
0 to 3	77	23.3	Bungalow	38	11.5
4 to 6	203	61.3
7 and above	51	15.4

4.2. Data Analysis and Results

PLS-SEM using the SmartPLS 3 application enables the evaluation of the reliability and validity of the measurement model [40]. PLS-SEM assesses the measurement and structural models simultaneously, thereby running factor analysis to assess the convergent and discriminant validity and hypothesis testing at the same time [41]. The analysis of the measurement model involves four stages: (1) the individual reliability of the indicators, (2) the reliability of the constructs, (3) convergent validity and (4) discriminant validity. Table 4 shows that all loadings except Risk1, Risk2, Risk4, Anticipated Affect 4 and Anticipated Affect 5 are above the standard values for the requirement of variance extracted (AVE) and composite reliability (CR) and thus demonstrate that the measurement model are appropriate and fit [42]. Furthermore, this study’s Value Inflation Factor (VIF) was less than 3.3 which denotes no collinearity issues faced by the study. The results of discriminant validity test which comprises Fornell–Larcker criterion and HTMT ratio were presented in the Table 5. The results suggest that all constructs have high degree of constructs which show that the measurement model has no issue in terms of the discriminant validity [43].

Table 4. SmartPLS Measurement model of the study.

Construct	Item	Loadings	CR	AVE
Trust	Trust 1	0.928	0.957	0.849
	Trust 2	0.919
	Trust 3	0.901
	Trust 4	0.937
Cost	Cost 1	Deleted	0.889	0.799
	Cost 2	0.906
	Cost 3	0.882
Benefit	Benefit1	0.905	0.95	0.761
	Benefit2	0.901
	Benefit3	0.916
	Benefit4	0.886
	Benefit5	Deleted
	Benefit6	0.837
	Benefit7	0.777
Risk	Risk1	0.639	0.846	0.53
	Risk2	0.638
	Risk3	0.858
	Risk4	0.601
	Risk5	0.861
Anticipated Affect	AnticipatedAffect1	0.954	0.772	0.512
	AnticipatedAffect2	0.959
	AnticipatedAffect3	Deleted
	AnticipatedAffect4	0.339
	AnticipatedAffect5	0.318
Attitude	Attitude1	0.85	0.918	0.691
	Attitude2	Deleted
	Attitude3	0.884
	Attitude4	0.849
	Attitude5	0.786
	Attitude6	Deleted
	Attitude7	0.781
Social Norm	SocialNorm1	0.933	0.921	0.854
	SocialNorm2	0.915
	SocialNorm3	Deleted
SEPBC	SEPBC1	Deleted	0.887	0.797
	SEPBC2	0.882
	SEPBC3	0.903
Intention	Intention1	0.928	0.925	0.861
	Intention2	Deleted
	Intention3	0.927

Note: Cost1, Benefit 5, Anticipated Affect 3, Attitude 2, Attitude 6, Social Norm3, SEPBC1, and Intention2 were deleted due to low loading.

Table 5. SmartPLS Discriminant validity of the study.

Fornell–Larcker Criterion										Heterotrait–Monotrait Ratio (HTMT)
	AA	ATT	BE	CO	INT	RI	SE	SN	TR		AA	ATT	BE	CO	INT	RI	SE	SN	TR
AA	0.713									AA
ATT	0.661	0.831								ATT	0.638
BE	0.649	0.722	0.873							BE	0.632	0.788
CO	0.592	0.614	0.545	0.894						CO	0.403	0.358	0.243
INT	0.573	0.612	0.468	0.535	0.928					INT	0.509	0.622	0.524	0.678
RI	0.492	0.537	0.483	0.516	0.651	0.743				RI	0.259	0.128	0.174	0.464	0.207
SE	0.314	0.394	0.313	0.378	0.549	0.171	0.893			SE	0.354	0.48	0.367	0.696	0.672	0.357
SN	0.226	0.294	0.208	0.365	0.532	0.081	0.504	0.924		SN	0.443	0.714	0.545	0.48	0.657	0.142	0.639
TR	0.013	0.113	0.054	0.151	0.051	0.089	0.473	0.534	0.921	TR	0.522	0.67	0.578	0.434	0.733	0.112	0.563	0.604

Note: AA: Anticipated Affect; ATT: Attitude; BE: Benefit; CO: Cost; INT: Intention; RI: Risk; SE: Self Efficacy Perceived Behavioral Control; SN: Social Norms; TR: Trust.

4.3. Structural Model Evaluation

The study used a structural model method to test the developed hypotheses in the second step of SmartPLS. Analyzing the R-Square (R2), Variance Inflation Factor (VIF), multicollinearity, effect size (F2), model fit, coefficients, p-values, and t-values are integral parts in the structural model evaluation [43]. In this study, Table 6 presents these results. The R2 value is above 0.35 [44] for Intention, Attitude and Anticipated Affect showing that it is an appropriate model. Figure 2 sets out the path model with R2 value. The effect size (F2), indicates the evaluation of the change in R2 and the relevance of the explanatory variable in justifying the measured variable. The effect size of Attitude on Intention was 0.085 (minor effect), Social Norm on Intention was 01.26 (minor effect), SEPBC on Intention was 0.047 (minor effect), Cost on Attitude was 0.027 (minor effect), Risk on Attitude was 0.014 (minor effect), Benefit on Attitude was 0.341 (significant effect), Anticipated Affect on Attitude was 0.149 (moderate effect), Trust on Cost was 0.153 (moderate effect), Trust on Risk was 0.008 (minor effect), Trust on Benefit was 0.422 (significant effect), and Trust on Anticipated Affect was 0.539 (significant effect) [43,45]. All endogenous variables had Q2 values greater than zero, indicating that the model was predictive. In PLS-SEM, the standardized root means square residual (SRMR) was also used to assess model fit [43]. The SRMR value for this study is 0.03, indicating a good model fit.

Table 6. SmartPLS Structural model of the study.

	Relationship	Std Beta	Std Error	t-Value	Decision	R2	F2	Q2
H1a	Trust → Cost	0.365	0.062	5.876 **	Supported	0.133	0.153	0.101
H1b	Trust → Risk	0.089	0.092	0.968	Not Supported	0.008	0.008	0.001
H1c	Trust → Benefit	0.545	0.062	8.735 **	Supported	0.297	0.422	0.222
H1d	Trust → Anticipated Affect	0.592	0.055	10.779 **	Supported	0.35	0.539	0.163
H2a	Cost → Attitude	0.109	0.038	2.858 *	Supported	0.599	0.027	0.407
H2b	Risk →Attitude	0.075	0.052	1.425	Not Supported	0.599	0.014	0.407
H2c	Benefit → Attitude	0.486	0.053	9.153 **	Supported	0.599	0.341	0.407
H2d	Anticipated Affect → Attitude	0.322	0.045	7.177 **	Supported	0.599	0.149	0.407
H3	Attitude → Intention	0.28	0.059	4.726 **	Supported	0.431	0.085	0.367
H4	Social Norm → Intention	0.221	0.067	3.293 **	Supported	0.431	0.126	0.367
H5	SEPBC → Intention	0.311	0.05	6.158 **	Supported	0.431	0.047	0.367

Note: * p < 0.05; ** p < 0.01.

4.4. Hypotheses Testing

Figure 3 and Table 6 show the findings of the hypotheses testing. The results from the PLS-SEM show that H1a (β = 0.365, p < 0.01), H1c (β = 0.545, p < 0.01) and H1d (β = 0.592, p < 0.01) were supported, while H1b was rejected. In terms of the relationship between cost, risk, benefit and anticipated affect on attitude, the results show that H2a (β = 0.109, p < 0.05), H2c (β = 0.486, p < 0.01) and H2d (β = 0.322, p < 0.01) were supported, while H2b was rejected. For the relationship between attitude and intention, the result shows a positive significant effect (β = 0.280, p < 0.01), thereby supporting H3. Additionally, there is a significant positive effect on social norm and intention (β = 0.221, p < 0.01) which supports H4. Finally, the relationship between SEPBC on intention has a significant relationship (β = 0.311, p < 0.01), thus supporting H5.

Figure 3. SmartPLS Path Model.

5. Results Discussion

This study covers the three paths of TAF by explaining the relationships within consumer acceptance of BESS including: (1) trust, (2) the combined effects of affect and cognition, (3) consumer’s outcome evaluation (attitude and subjective norms) and perceived behavioral control and (4) intention to accept the BESS technology. The results confirmed the significant relationship between the three paths of TAF in explaining the consumer acceptance of BESS. The top three most significant relationships include the relationships between trust and anticipated affect (t value = 10.779), benefit and attitude (t value = 9.153), and trust and benefit (t value = 8.735). Secondly, it shows that trust has a positive significant relationship with the consumer’s perceived cost, perceived benefits and anticipated affect. The findings are consistent with the study by [22] which found that trust in a municipality or industry has a positive impact on individual perceived effects and anticipated affect on public acceptance of BESS in Germany. On the other hand, there was no significant relationship found between trust and perceived risk. This could be due to the consumers’ lack of knowledge of the risks associated with BESS [24]. Thirdly, it demonstrates that the consumers’ combined effects of affect and cognition such as perceived cost, perceived benefit and anticipated affect have a significant effect on consumer attitude towards BESS. The findings are consistent with the study by [24] which proved that anticipated affect influenced the attitude towards BESS. Lastly, this study demonstrated that the consumer’s attitude, subjective norms and self-efficacy or perceived behavioral control have a positive significant relationship with their intention to use BESS in the future. The findings are consistent with previous studies by [24] in Canada and [26] in the United Kingdom. This study contributes to the theory of Technology Acceptance Framework (TAF) by confirming the linkages between consumer’s trust in BESS; the combined effects of affect and cognition (the perceived cost, perceived risk, perceived benefit and anticipated affect) on BESS; consumers’ combined effects of affect and cognition (perceived cost, perceived 185 risk, perceived benefit and anticipated effect) of BESS and their attitude towards BESS.

6. Conclusions and Future Works

The study highlight several implications for understanding the residential consumers’ acceptance of BESS, particularly in a developing country such as Malaysia. Firstly, the findings proved the important role of trust in formulating the consumer’s perceived benefits, perceived cost and anticipated affect of BESS. The issue of trust in utilities, BESS service providers and policy makers is a very important factor in the acceptance of BESS by residential consumers BESS [22]. Secondly, given the insignificant relationship between consumer’s perceived risk and their intention to accept BESS, the administrative bodies must educate the public about the safety aspect of BESS, the impact of BESS on the environment, battery recycling and reuse risks and financial risks associated with BESS [23]. Thirdly, the findings of this study provide insights for policy makers when designing policies related to BESS since the acceptance by the consumer is likely to be contingent on the trust, perceived costs, perceived benefits, perceived risks, attitude, social norms and self-efficacy or perceived behavioral control of the consumer [46]. Lastly, the findings regarding consumer acceptance will provide an insight for those designing a suitable business model for BESS, whether for wholesale ownership, substation ownership and/or end-use customer ownership [44].

This study offers a foundation for future empirical studies related to the acceptance of green energy technology. Moreover, the replication of the study in a different country in a similar region, such as Southeast Asia, can provide insights on the acceptance of BESS. On top of that, as the study has been undertaken to collect data from a cross-sectional study, future work on the acceptance of the BESS can focus on empirical work based on longitudinal data. Lastly, future work can access a larger sample drawn exclusively from commercial and industrial consumers instead of focusing on residential customers since this type of consumer consumes a large amount of electricity.