Consumer-Centric Rate Design for Peak Time Energy Demand Coincidence Reduction at Domestic Sector Level-A Smart Energy Service for Residential Demand Response

Reducing demand coincidence of customers with distribution system peak hours is very essential in the modern world energy sector. There is a great scope for peak demand coincidence reduction at the customer level, especially in the residential sector through Residential Demand Response programs. Through smart meter installations along with IT-enabled technologies, many of the distribution company’s initiatives like residential demand response programs can be taken to domestic customers with ease, less cost, and less technology deployment. Rate design is one such effective approach. Even though time-varying rates like Time of Use have been used as an effective approach for reducing peak electricity demand in different sectors around the world, the residential sector has not gained much attention due to a few challenges like externality problems, high on-peak and low off-peak prices, improper pricing mechanisms. Hence, considering the above challenges and constraints, a rate design, named consumer-centric time of use tariff is proposed in this article for domestic customers. The tariff is consumer-centric such that each customer gets a unique on-peak unit price and off-peak unit price based on their consumption during peak hours both at the house-level and utility level rather than common and fixed on-peak and off-peak prices for all the customers, thus addressing the above-mentioned constraints. For this, customers are classified into different clusters using the Machine Learning Algorithm K-Means. The proposed rate design model has been analyzed on synthetic smart meter data of 10 houses, and it is observed that the proposed tariff shows an increase in the monthly revenue by 4.3% for the utility and a variation of -0.4% to 7% of energy charge for different customers. This study and analysis show that the proposed consumer-centric time of use rate design provides a better pricing mechanism with a win-win strategy for both customers and utility, thereby avoiding windfall gains or losses to both. Furthermore, the proposed tariff influences each residential customer of different consumption levels to reduce peak demand coincidence as well as energy consumption for the power sector.


Introduction
Peak load or peak demand is one of the ve smart gridspeci c drivers in the electricity sector identi ed by the International Energy Agency (IEA), with issues such as signi cant stress on the grid, increasing the risk of blackouts and brownouts, increased power prices for consumers, and the need to build additional plants. To meet this peak demand due to uctuating load variations, such as increased demand in the early morning or evening, grids must provide additional peaking capacity, i.e., additional energy generating capacity, which increases the utility's cost. Furthermore, smart metering and proper electricity rate design for consumers must be developed as part of Demand Response (DR) programs for peak load management. Small customers, such as Low Voltage (LV) customers, have been subsidized by large customers, such as HV customers, and low load factor customers have been subsidized by high load factor customers under volumetric rates. Any change in rate design that corrects these subsidies will result in lower bills for overpaying customers and higher bills for underpaying customers [1]. e process of setting prices and price structures to be paid by customers with a fair allocation of costs across customer classes so that utilities have a reasonable opportunity to recover their revenue requirement is known as rate design (costs plus allowed return). Applications of Electrical Engineering (AEE) believes that of the various rate design options available for mass-market customers (residential and small commercial customers), the evolution towards time varying rates (TVRs), that is price electricity based on granular and precise price signals of kWh usage and/or kW demand, is the preferred long-term option for modern rate design [2]. Time-varying pricing is when prices change over time, and different prices apply at different times on different days [3]. TVRs generally work by charging more for electricity when demand on the system is high. TVRs give customers more control over their electricity bills and usage while also assisting utilities in avoiding investments in additional energy infrastructure, such as polluting power plants and transmission lines, built to meet rising electricity demand. A 210-megawatt power plant can be avoided by adopting TVR by just 20% of customers, according to an Oklahoma utility.
TVR is not a novel idea. TVR already has many large commercial and industrial customers all over the world. However, it is less common among residential customers.
ere is now a lot of room for peak demand management programmes at the residential level in India, thanks to enabling technologies like advanced metering and services that provide customers with the information and tools they need to respond to price signals. TVRs are classified into four broad categories, as shown in Figure 1, Time-of-Use (ToU), Critical Peak Pricing (CPP), Peak Time Rebate (PTR), and Real Time Pricing (RTP). TOU is the most basic pricing scheme, with a fixed pricing structure for each of the day's predefined peak and off-peak time periods. RTP charges hourly rates based on day-ahead market prices or real-time spot market prices for electricity and is the most variable and sophisticated. In between are CPP, which charges consumers fixed high peak prices only on certain days of critical peak demand that are announced in advance, and PTR, which offers rebates to customers based on their peak time consumption reduction during a peak demand event.
ToU has gained popularity among TVR types (refer to Figure 1) due to benefits such as preestablished prices for consumers that are easy to understand and typically do not change more than twice a year. Aside from these benefits, existing ToU pricing schemes reveal some constraints and challenges.

TOU Pricing Benefits and Challenges in Different
Countries. ToU encourages electricity use reduction during peak demand periods by shifting consumption to periods of low demand by offering a lower price during these periods and a higher price during peak demand periods. e potential benefits of ToU, as identified by various pilot studies, include noticeable peak load reduction [5], avoided investments in additional energy infrastructure [6], reduction in customers' energy bills [7], distributed generation deployment with improved retail pricing fairness [8], environmental benefits [9], and lower wholesale market prices [10]. ToU rates as the most widely used Demand Response strategy in the United States, with over twenty lakh customers [11,12]. e use of ToU rates as a demand response strategy in various European countries, including the United Kingdom, Italy, France, and Spain, as well as in other countries such as China, is discussed in [13][14][15]. A reduction in peak use of 3% to 6% was observed from real world experiments on ToU rates [16]. e successful results of a survey of 43 ToU pricing programmes in the industrial sector show that cost savings ranging from 72.0% to +82.6% can be achieved depending on different switching strategies and specific utility programmes [17]. Low-income households, on the other hand, had mixed feelings about traditional ToU pricing, such as lifestyle curtailment [18] and financial burden, if low-income customers are unable to shift peak consumption [19,20]. e authors of [21] discovered that low-income households are price responsive, resulting in a 13% reduction in average peak demand in a ToU pricing experiment. Furthermore, the study in [22] claims that with appropriate design and ToU pricing scheme selection, low-income consumers can achieve a good amount of peak load shift, whereas a detailed analysis in [23] shows that the higher a household's income, the lower the response to the price signal. Despite the fact that model experiments in [24,25] show a peak use decrease from 50% to 7.6%, consumers' costs have increased without much benefit due to higher (average) prices and higher electricity use as a result of lower tariffs during the off-peak period. ere is a possibility of a new peak demand with the ToU tariff at the start of off-peak hours due to customers scheduling appliances due to the price drop right after on-peak [26]. Because ToU rates are not dynamic and are not based on actual wholesale market prices, they may be less useful in specific grid events. However, the ToU raised awareness about energy consumption, which resulted in an overall energy reduction following the ToU's implementation. According to the above literature review, the following challenges and constraints of ToU pricing have been identified: (1) A few ToU models that charge a single fixed price to all customers of different consumption levels for a specific period affect low-level customers.
(2) A few ToU tariffs charge customers higher average prices, resulting in benefits to utilities and cost burdens to customers rather than benefits to customers. (3) Offering much lower tariffs during off-peak periods may result in higher electricity use at lower prices by a few classes of customers, resulting in revenue loss for the utility but being beneficial to high-use customers. (4) e ToU pricing scheme may result in an externality problem, which has the effect of high-level customer consumption on the price rates of other customers, particularly during peak periods.
In order to address the aforementioned constraints and challenges, this article proposes a TOU tariff in which each customer of different consumption levels receives individual on-peak and off-peak unit prices rather than a common single fixed price for all customers based on their monthly average unit price. As a result, it focuses on each individual customer to reduce individual on-peak/off-peak ratios, resulting in a lower peak load at the utility level and benefits for each customer. As a result, the proposed tariff benefits both customers and the utility.

Scope of TOU Pricing Implementation in India's Electric
Domestic Sector. India was chosen as a case study because it is a large country with a large population that has made tremendous progress in the power sector by increasing electricity access to its citizens in recent years. India's energy market has enormous potential as the country's power sector transitions to smart technology. In recent years, India's power sector has made significant strides in generation and transmission. However, distribution inefficiencies such as rising costs and consumer tariff challenges, as well as high aggregate technical and commercial (AT&C) losses, have hampered distribution utilities' future growth. Over the last decade, India's power sector has successfully added massive new power generation capacity to meet rising energy demand, reducing the country's energy deficit from a massive 8.5% in 2010 to just 0.5% in 2019-20. Peak power demand deficits have also decreased dramatically, from 9.8% in 2010 to 0.7% in 2019, according to data published by the Central Electricity Authority (CEA) [27]. Electricity generation in India to meet energy demand varies by state, depending on population, occupation (domestic, agricultural, commercial, and industrial), temperature, seasonality, and cultural practices. Differential pricing at peak periods in the Indian power sector would likely result in efficiency gains by taking into account different types of energy demand, supply variations, and service costs [28]. e Institute for Energy Economics and Finance Analysis (IEEFA) recently published a report highlighting the importance of introducing hourly rates and day-ahead market pricing for the Indian electricity sector in order to reduce peak power demand and encourage usage when energy demand is typically low [29]. However, the current pricing system in India is primarily based on flat-rate tariffs and slab-based tariffs, providing no incentives to customers to manage peak load. e residential sector, which consumes the most electricity in India, has a lot of room for peak demand management with Time Varying Rates like Time of Use (TOU). According to the World Bank, domestic electricity consumption will have increased 260% by 2021 [30]. Many high-use residential consumers contribute to high peak demand by using home appliances such as air conditioning and heating during peak hours. According to the International Energy Agency's (IEA) World Energy Outlook 2018, the number of Indian households with air conditioners (ACs) has increased by 50% in the last five years. Two-thirds of Indian households are expected to own an air conditioner by 2040, a staggering 15-fold increase from today. Most domestic consumers are unaware of the impact of peakperiod electricity use on electric network costs and the environment. In addition, a lack of education programmes and incentives for consumers to shift their electricity use from peak to off-peak periods has an effect on-peak energy consumption in the residential sector. Individual consumer preferences influence energy consumption in India's domestic electricity sector. At the moment, most Indian states' tariff structures for domestic customers are slab-based, with little room for reduction in residential consumers' electricity bills other than lowering their monthly consumption. However, with time-varying prices such as TOU, there is a great opportunity for domestic customers to save money simply by changing their consumption patterns rather than reducing consumption. e TOU tariff is popular among High Voltage (HV) and Extra High Voltage (EHV) Journal of Electrical and Computer Engineering customers in many Indian states, including Maharashtra, Karnataka, Andhra Pradesh, West Bengal, Goa, Gujarat, Uttar Pradesh, and Kerala. In fact, in some states, TOUs are required for this group of customers. However, it is a novel concept for Indian residential consumers. e same TOU tariff structure as for HV cannot be applied to the domestic sector because most states in India use slabbased tariffs for residential customers and flat-rate tariffs for HV customers.
e research presented in [31,32] provides compelling evidence that TOU pricing can reduce peak electricity demand and increase energy conservation in the residential sector. Optimal dayahead self-scheduling and operation of prosumer microgrids with hybrid machine learning-based weather and load forecasting have been reported [33]. A new collaborative framework for fair and cost-effective resource allocation in a low-voltage electricity community is well presented [34]. Creating an energy management system for prosumers based on a modified data-driven weather forecasting method is well reported [35].
According to the discussion above, there are still promising opportunities to improve India's electricity distribution sector, particularly by involving residential sector end users. As a result, the goal of this research is to propose a new cost-reflective electricity pricing scheme and a consumer-centric TOU (C-TOU) rate design for domestic customers in India, which encourages each individual residential customer to manage the country's growing peak demand. e proposal seeks an alternative solution to the above-mentioned constraints on existing TOU tariffs.

Research Highlights.
(1) To the best of our knowledge, the proposed C-ToU tariff design as an RDR strategy is a novel approach that focuses on each customer to reduce their consumption during peak hours both at the house level and utility level.
(2) e proposed tariff is consumer-centric such that each customer of different consumption levels gets a unique on-peak unit price and off-peak unit price based on their monthly consumption during peak hours both at house level and utility level rather than common and fixed on-peak and off-peak prices for all the customers or cluster of customers, thus finding a solution to the externality problem as discussed above. (3) e tariff is so designed that the average unit with the proposed C-ToU does not vary much from the present base tariff, thus resulting in the win-win strategy for both customers and utility and thereby avoiding windfall gains or losses to both. (4) Peak and off-peak prices are designed in such a way that discourages the higher electricity use by a few classes of customers due to lower prices during offpeak periods, thus resulting in revenue loss for the utility.
(5) Customers are classified into 4 clusters based on their consumption during peak hours at their house level and utility level, and then, charges are determined for each customer rather than for a cluster that readdresses the challenge of overpaying customers' bills to go down and underpaying customers' bills to go up. (6) e single model helps to identify the high contributors to peak time consumption at the utility level, as well as at house level through clustering and hence helps to implement RDR programmes easily. (7) us, the model encourages each customer to reduce their consumption during peak hours to reduce the demand coincidence with utility during peak hours.

Proposed Rate Design: Consumer-Centric ToU (C-ToU)
e term "consumer-centric" is used in this work to indicate that the proposed ToU price is oriented toward individual domestic customers, with each customer charged with a different on-peak and off-peak unit price based on their monthly average unit price rather than a common fixed price for all that is easy to understand and make corresponding changes to reduce the monthly bill. Consumer-centric ToU (C-ToU) pricing is a proposed tariff that combines twoperiod ToU pricing rates (peak and off-peak) with a base tariff. is tariff is designed primarily by taking each customer's energy charge into account and does not take into account other costs such as fixed charges and wheeling charges, which are utility dependent and cannot be changed by the customer. According to market research, three factors influence time of use rates.

e Price Ratio between On-Peak and Off-Peak Period.
Most existing TOU tariffs or proposals in various countries charge a predetermined fixed on-peak unit price and offpeak unit price that is common for all customers or customers in a specific block, which may result in externality issues, primarily for low-level consumers. In general, utilities charge all customers a peak tariff that is 20%-30% higher than the normal tariff per kWh and an off-peak tariff that is 15%-20% lower than the normal tariff per kWh.
is common tariff may result in a financial burden for low-level customers or a loss of utility revenue due to increased electricity use at a low cost by a few classes of customers as a result of lower tariffs during the off-peak period. e results in [36] state that to achieve the effectiveness of TOU pricing schemes, a significant price difference between the peak and off-peak hours is required. In the present work, a common price ratio is not considered for all the customers, which may lead to injustice to others. Instead, a price ratio is considered based on the total monthly consumption of each customer in such a way that the off-peak price rate increases with the increase in the total monthly consumption of each customer (more the total monthly consumption, the more the off-peak price). For example, if a customer with 300 kWh consumption per month and a customer with 700 kWh consumption cannot be charged with the same on-peak and off-peak price, which may lead to injustice, resulting in a very low bill compared to the present tariff.
Considering this issue, a novel TOU tariff, C-TOU, is proposed in such a way that each customer will get a unique on-peak and off-peak unit price based on their monthly consumption and monthly average unit price with a base tariff rather than a common on-peak and off-peak unit price for all. In this proposed tariff, on-peak unit prices are charged at 20% higher than the monthly average unit price of an individual customer, and the off-peak price is set at a level that helps to recover the revenue amount fully from the base rate. us the off-peak price is charged at 10%-20% lower than the monthly average unit price of individual customers based on the class of the customer, instead of the common on-peak and off-peak unit prices for all the customers. For this, residential consumers are classified into different classes based on their monthly consumption during peak hours, as recorded in Table 1.
In Table 1, House-level monthly peak ratio gives the ratio of kWh usage during peak hours to the total monthly kWh usage of that particular house.
House − level monthly peak ratio � kWh usage during peak hours per month total kWh usage per month , whereas utility-level monthly peak ratio gives the ratio of kWh usage during peak hours of a particular house to the monthly total peak hours kWh fed by the utility.
Utility − level monthly peak ratio � kWh usage during peak hours of a particular house per month total kWh fed by the utility during peak hours per month . (2) us, house-level monthly peak ratio helps to analyse a consumer's contribution to peak time consumption at his house level, whereas the utility-level monthly peak ratio helps to analyse a consumer's contribution to peak time consumption at the utility level. Based on these 2 factors, on-peak rate coefficient, the rate at which onpeak unit price (charge for consumption during peak period) is offered to the customer and off-peak rate coefficient, the rate at which off-peak unit price (charge for consumption during off-peak period) is offered to the customer can be determined.
In designing the above rate coefficients, first the on-peak rate coefficient is set as a certain percentage (120%) of the base tariff average unit price. en the off-peak rate coefficient is set at a level that does not cause much variation in the monthly average unit prices and monthly bill from the present base tariff for each customer class mentioned above and fully recovers the revenue. us, the on-peak and offpeak unit price for each customer are calculated as shown in individual on − peak unit price � (on − peak rate coefficient) ×(individual monthly average unit price), individual off − peak unit price � (off − peak rate coefficient) ×(individual monthly average unit price), (4) e 50th percentile of the house-level monthly peak ratio and the utility-level monthly peak ratio of all houses is considered as the baseline energy consumption. e 50th percentile is the median of a distribution of data and is the point in the data where 50% of the data falls below that point and 50% falls above it. is classification helps to identify consumers who are contributing more to peak time consumption.

e Length of the Peak Period.
e length of the peak period should be neither too short nor too long for the domestic sector, as a longer peak period may not attract a customer, whereas shorter peak periods may not be beneficial for the utility. Residential peak load varies based on the number of people at home. Many Indian utilities face early morning residential peak loads most of the time in a year with people preparing for school or work. Once the residential customers are away from home during the day, this power usage drops. Again, demand reaches its peak later in the evening when most of the customers are back home, with high usage of electrical appliances like ACs, TVs, lights, and fans, typically between 6 p.m. and 9 p.m. However, in summer, late morning or afternoon peak loads along with evening peak loads are experienced by many Indian utilities due to cooling loads like air conditioning, coolers, and fans that contribute to the highest consumption of electricity. In the present work, peak periods are proposed and considered for the analysis based on typical load curve Journal of Electrical and Computer Engineering illustrating the national variation in demand (aggregated electrical load of the different states in India) over a specific time period as shown in Figure 2. e climate of India is broadly classified into two seasons for the purposes of considering seasonal variations and demand patterns in this study: summer season (March-June) and nonsummer season (July-February). Taking these factors into account, the following on-peak and off-peak periods are proposed for India's residential sector (refer Figures 3 and 4).
Sundays are considered off-peak. Public holidays are considered normal weekdays. us, a peak period length per day of 6 hours for nonsummer months and 9 hours for summer months is proposed and considered for the analysis in the present work. However, as the geographical area of India is quite large, considering a common peak and off-peak time and a common tariff for all the states may not reap the same benefits. Hence, based on local seasonal conditions, different states may consider their peak hours. To maintain the financial balance at the utility level, an inverse relationship between the length of the peak period and the off-peak price needs to be maintained such that lesser the peak hours, the more the off-peak price.

e Number of Pricing Periods.
Proposed consumercentric ToU (C-ToU) pricing uses two-period ToU pricing rates (peak and off-peak) along with the present slab-wise tariff (state wise) as there is a possibility of many customers losing control over their bills with a greater number of pricing periods.

Methodology
Consumer-centric ToU pricing (C-ToU) is a combination of two-period ToU pricing rates (peak and off-peak) and a base tariff.
is tariff is designed primarily by taking each customer's energy charge into account and does not take into account other costs such as fixed charges and wheeling charges, which are utility dependent and cannot be changed by the customer. is model was created with India as a reference, where smart meter installations are currently underway. Figure 5 depicts the proposed C-ToU tariff methodology. According to the C-ToU model, the energy charge is the sum of on-peak and off-peak energy charges.
where n � number of days in a month, j � time period of the day, EC IM represents individual monthly energy consumption, and EC represents energy consumption.  Table 2, total energy charge for a month can be calculated as follows.

Calculation of Average Unit Price from Monthly Energy Charge and Number of Consumed Units.
Each customer's monthly average unit price is calculated based on present residential tariff in that particular state by using individual average unit price � individual monthly energy charge(Rs) individual monthly energy consumption(kWh) , where individual monthly energy charge is calculated from the base tariff structure and individual monthly energy consumption is calculated for each consumer by summing the half-hourly consumption for all the days in a month.
23: 00 j�00: 00 where n � number of days in a month, j � time period of the day, EC IM represents individual monthly energy consumption, and EC represents energy consumption.

Stage 3.
To separate each customer's monthly on-peak and off-peak units based on timings, kWh usage during peak and offpeak times can be easily separated using smart meter data based on ToU timings. Total monthly on-peak and off-peak energy consumption is calculated by summing half-hourly consumption data during the specified ToU hours mentioned in Figure 3 and Figure 4 for all days in a month based on the peak and offpeak timings discussed in Section 2 of this work. Individual monthly on-peak energy consumption (EC IMon-peak ) for a nonsummer month is as follows: where n � number of days in a month of a nonsummer month and j � specific time period of the day. Individual monthly on-peak energy consumption (EC IMon-peak ) for a summer month is as follows: where n � number of days in a month of a summer month and j � specific time period of the day. Off-peak units are obtained by removing the on-peak units from the total consumption as shown in the following.
Individual monthly off-peak energy consumption is as follows:

Stage 4
3.4.1. Determining the Class of the Customer. Once stage 1, stage 2, and stage 3 are obtained, on-peak and off-peak energy charges must be obtained for which rate coefficients obtained using the criteria shown in Table 1. Customers are divided into different sections using the Machine Learning Algorithm and K-means, based on the criteria shown in Figure 6. K-means clustering is a well-known and highly effective unsupervised machine-learning algorithm [38]. is algorithm generates clusters by grouping similar items into clusters. K represents the number of formed groups.
Because real smart meter data are not available due to the pilot stage of smart metres, K-means clustering is performed on synthetic smart meter data from 10 houses in this work. ese data were generated using synthetic data of consumption on a half-hourly basis for all days of the year based on each customer's home appliances, daily usage pattern, and lifestyle as obtained through personal communication. ese ten houses are thought to be from Andhra Pradesh, a southern Indian state (AP). e study considers various types of customers, such as a middle-class family of four members, a large family of six members with a lavish lifestyle, four working bachelors sharing a flat, self-employed at homes such as a beauty parlour with three members in a family, both members of the family working, a house with four members and comfortable appliances such as a coffee maker, toaster, home theatre, clothes dryer, gaming PC, and upper and lower middle-class families. ese customers use LED and CFL bulbs, water-pump motors (370 W, 746 W), electric-irons, ovens, grinders, washing machines (550 W, 750 W), televisions (100 W, 200 W), refrigerators (250 W, 350 W), air-coolers, kitchen chimneys, room heaters (2000 W), rice cookers (500 W, 1200 W), exhaust fans, hair dryers, dishwashers, grinders, geysers, treadmills, and vacuum cleaner. Some of these customers had two refrigerators, televisions, and air conditioners. Figures 6 and 7 show average hourly load profiles for each customer for one summer and one nonsummer month.
Customers in cluster 1 are considered "critical" customers according to Table 1 and Figure 8 because they have high usage both at home and at the utility level during peak hours. Cluster 2 customers are those who have lower peakhour consumption at home but contribute significantly to peak-hour consumption at the utility level. As a result, they are classified as "high." Customers in Cluster 3 contribute less to the utility level, but their consumption during peak hours exceeds 50% of their monthly kWh usage. As a result, they are also classified as "high." Customers in Cluster 4 are considered "moderate" as their contribution at the utility level is less than 50%, and their peak time consumption at the house level is less than 50% of their total consumption. Figure 9 depicts the same classification for a nonsummer month.

Determining On-Peak and Off-Peak Rate Coefficients
Based on the Class of the Customer. Based on the clustering as shown in Figures 8 and 9 and criteria as recorded in Table 1, rate coefficients are obtained for each customer.
us, individual monthly energy charge can be obtained as shown in equations (14) and (15) For nonsummer (14), For summer (15), UP Ion− peak and UP Ioff− peak can be obtained from equations (3) and (4).

Results and Discussion
e proposed C-ToU electricity pricing scheme is applied and analyzed on synthetic consumption data created for 10 houses on a half-hourly basis for all days of the year based on each customer's home appliances, daily usage pattern, and lifestyle obtained through personal communication. Monthly energy consumption, monthly total on-peak consumption, off-peak consumption, monthly energy charge-with current base tariff and proposed tariff, difference between the tariffs, house-level monthly peak ratio%, and utility-level monthly peak ratio% for a summer and nonsummer month are shown in Tables 3 and 4. It can be observed from the "Diff" column in Table 3 that houses H2 and H4 can pay less with the proposed tariff due to their lower peak hour consumption but are paying high with the present base tariff at state level, whereas the rest of the houses need to pay more to the utility due to their high peak consumption but are paying less under the present tariff. Similarly, Table 4 shows that in a nonsummer month, houses H2, H3, and H5 can reap profits, whereas other houses need to pay more to the utility. Figure 10(a) shows the % monthly consumption during peak and off-peak hours at the house level of each house, and Figure 10(b) shows the % monthly consumption during peak time at the utility level of each house for a summer month. Figures 11(a) and 11(b) show the same for a nonsummer month. Figure 12 shows the difference in monthly energy charges experienced by each customer with proposed C-ToU compared to the base tariff for a summer month and a nonsummer month.

Individual Customer Level.
us, it can be observed that customers with low peak time consumption can reap good monthly savings without any modifications, whereas customers with high consumption during peak hours are charged accordingly with the C-ToU tariff. Customers H6, H8, H9, and H10 who are in the "critical" cluster experience a "noticeable" energy charge variation with the proposed C-ToU from the base tariff. ese customers are considered "treatable" as their contribution to the peak at utility level and house level is very high. Houses H5 and H7 that are in cluster "high1" customers experience "considerable" energy charge  Utility Level monthly peak ratio (%) Hours level monthly peak ration (%) Figure 9: Customer classification based on monthly house-level and utility-level peak time consumption in a nonsummer month. energy charge variation with the proposed C-ToU from the base tariff. ese customers are considered "encourageable" as their contribution to the peak at utility level and individual level is moderate. Figure 12(b) shows the same for a winter month. . Figures 13(a) and 13(b) show the comparison of monthly revenue in the form of energy charges that can be collected by the utility with C-ToU and base tariff for summer and nonsummer months.

Utility Level
It can be observed that with the same consumption, C-ToU yields increased revenue collection compared to the base tariff at the utility level. is proposed tariff results in marginal profit to the utility compared to the present tariff as well as peak-time energy consumption. erefore, from the above analysis, it is evident that the proposed C-TOU tariff plan provides a better pricing mechanism that is consumercentric and thus influences each residential customer of different consumption levels to reduce peak demand as well as overall energy consumption for India's electricity sector [39].

Conclusion
Even though TOU has been adopted as a DR strategy for residential customers in many countries, few constraints and challenges like externality problems, unjustified price ratios for different categories of customers, etc., have been wandering existing TOU schemes. Addressing above-mentioned constraints and the advantages of TVR pricing schemes like TOU for domestic customers, a C-TOU has been proposed in the present work as the RDR strategy. e proposed C-TOU charges unique on-peak and off-peak unit prices for each customer based on their monthly average unit price rather than a common price for a slab or a block of customers. For this, customers are classified into clusters using the ML algorithm K-means. Analysis performed on 10 houses synthetic data shows that the proposed C-TOU tariff plan provides a better pricing mechanism with low variation of monthly electricity bills from that of the present tariff for customers of different usage levels, thus avoiding windfall gains or losses to customers and utility and influences each residential customer of different consumption level to reduce peak demand as well as energy consumption.A revenue gain of 4.3% at utility level is observed for both summer and nonsummer months on data of 10 houses with C-ToU compared to base tariff. A variation of -0.4% to 7% of monthly energy charge at house level is observed at different houses for both the months with C-ToU compared to base tariff.
is work shows that the C-TOU tariff is more beneficial than the base tariff for both utility and customers. e proposed scheme could be considered by decisionmakers in tariff plans for residential customers in different countries along with India where smart meter installments are in the progressive stage. Further study is needed to investigate the impact of the proposed TOU pricing scheme on real-time data.

Disclosure
is was a research work of Swathi G, Department of Electrical and Electronics Engineering, Koneru Lakshmaiah Education Foundation, Vaddeswaram, Andhra Pradesh, India.
Data Availability e data are used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that they have no conflicts of interest.