**5. Conclusions**

This work has developed a simplified and practical approach to model residential electricity load profiles where the model can match the main city's characteristics with the representative pre-defined households profiles proportionally. The Wepro model is advantageous as an efficient approach to develop the residential electricity load profiles at the city level, especially where survey data, smart-meter data or any other local temporal profiles dataset are unavailable. The findings concur with some load profile studies from the similar climate profile which indicate Winter as the highest consumption share and illustrate either December or January is having the highest consumption share. The results of the selected typical days for all load profiles indicate that most weekend days have a higher consumption share than weekdays in the same season. Moreover, all the hourly average load profiles in a year and per season demonstrate the consistent curve shapes, demand peaks and the peak hours within the season and model, either using Wepro-LPG or Wepro-ALPG. In terms of the curve shape and daylight characteristics between the models, the hourly average in a year of Wepro-ALPG is preferred to be used because it also shows a high similarity with the shape of the standard Dutch household provided by NEDU or previously EDSN, although the Wepro-ALPG load profiles illustrate less morning fluctuations as a result of the few input profiles taken by the approach. In addition, in terms of the evening peak, the hourly average in a year of Wepro-LPG is preferable to be used, because it resembles the evening peak time of the Dutch household characteristics, where the evening peak takes place after dinner time, which concurs with a Dutch load profile study that the evening peak takes place after dinnertime when e.g., TV, dishwasher, etc., are on because within the average Dutch household, cooking is done using gas instead of electricity.

Moreover, our work contributes by evaluating the characteristics of residential electricity load profiles based on time variation analyses: seasonal analysis, monthly analysis, days analysis and hourly analysis. In addition, this method is applicable to model previous year, current year and future year, where for current year and future year are used city's projected numbers.

Furthermore, the few selected household profiles which are the representative of the city's profile may dominate the shape of the output profiles where all of input have represented the city's age group, labour force composition and gender share. Although the few selected profiles may dominantly influence the output profile, based on the results, they still resemble the Dutch average household profile and concur with the common peak demands characteristics. In addition, although the Wepro model depends on external household profile generators such as LPG and ALPG, the Wepro model is found to be more efficient in storage capacity and computational process of the residential sector's load profiles, given the number of households in the city that can represent the local profile.

In future work, it would be interesting to identify the potential of energy savings based on the generated load profiles using a relevant machine-learning technique. We also look forward to add more main input parameter to the model and compare with the case study's measured data. Further work might also be conducted to extend residential electricity temporal profiles into spatial profiles.

**Author Contributions:** The idea, method and analysis of the study is designed by A.K. A.K., P.D.K.M. and P.S.N. wrote the paper. P.D.K.M. performed the complex pre-processing and modelling tasks in Python. P.S.N. reviewed and proofread the paper. All authors have read and agreed to the published version of the manuscript.

**Funding:** The research described in this paper is being conducted as part of the ClairCity project, funded by the European Union's Horizon 2020 research and innovation programme under grant agreement No. 689289, and a PhD fellowship within the CITIES project at Denmark Technical University (DTU) funded by the Indonesia Endowment Fund for Education (LPDP) under Letter of Guarantee: Ref:S-1401/LPDP.3/2016. The CITIES project is funded by InnovationsFund Denmark under contract: 1305-00027B.

**Acknowledgments:** We acknowledge ClairCity partners within the Technical work package UWE–United Kingdom, TML-Belgium, UAVR-Portugal, Techne-Italy, NILU-Norway, PBL-The Netherlands, CBS-The Netherlands, and other partners for supplying the related datasets and other large-scale inputs. We also acknowledge Noah Pflugradt for developing and publishing the LPG, and Gerwin Hoogsteen, the University of Twente, The Netherlands for developing and publishing ALPG. Furthermore, we thank Elke Klaassen for sharing her publication and The Netherlands' load profile information.

**Conflicts of Interest:** The authors declare no conflict of interest.
