 Research
 Open Access
 Published:
Prediction of domestic appliances usage based on electrical consumption
Energy Informatics volume 1, Article number: 16 (2018)
Abstract
Forecasting or modeling the onoff times of domestic appliances has gained increasing attention in recent years. However, comparing currently published results is difficult due to the many different datasets and performance measures employed. In this paper, we evaluate the performance of three increasingly sophisticated approaches within a common framework on three datasets each spanning 2 years. The approaches forecast the future onoff times of the appliances for the next 24 h on an hourly basis, solely based on historic energy consumption data. The appliances investigated are driven by user behavior and consume a significant fraction of the household’s total electrical energy consumption. We find that for all algorithms the average area under curve (AUC) in the receiver operating characteristic (ROC) is in the range between 72% and 73%, i.e. indicating mediocre prediction quality. We conclude that historic consumption data alone is not sufficient for a good quality hourly forecast.
Introduction
Forecasting or modeling the expected onoff times of domestic appliances is motivated from two directions: (i) generation of electrical load profiles and (ii) learning and predicting user behavior. Artificial load profile generation (Pflugradt 2016) can be helpful if large numbers of profiles spanning extended durations are required because their collection typically involves arduous measurement campaigns. While the precise prediction of the switchon/off times is in this case not the main concern, it is an essential part in applications targeting demand response systems: Learning the usage pattern of appliances and therefore, knowledge of the user behavior is a vital input to optimally plan energy usage (Chrysopoulos et al. 2014; Holub and Sikora 2013). While different prediction approaches have been published (Chrysopoulos et al. 2014; Holub and Sikora 2013; Truong et al. 2013; Barbato et al. 2011), an outstanding matter in adequately addressing the forecast of domestic appliance usage is a comparison of the available approaches: Published results are difficult to compare because of diverse performance metrics, different predicted appliances and the large variety of employed datasets, either measured at different geographic locations or even simulated. It is therefore unclear how well a method generalizes (i) over extended time periods and (ii) to other datasets with different attributes such as appliances, number of inhabitants, user habits and behavior.
In this work, we compare published the approaches we are aware of (Chrysopoulos et al. 2014; Truong et al. 2013; Barbato et al. 2011) and extensions from these on three datasets measured over 2 years in households located in Switzerland, Canada and the UK. We implemented these approaches into a common framework and compare their fitness in predicting the usage patterns of the appliances. In doing so, we focused on appliances, whose usage is mainly driven by user behavior and whose switchon time is flexible. In the relevant literature such appliances are commonly referred to as “shiftable loads”. Examples for such loads are washing machine, dish washer or tumble dryer. The Python source code for the experiments can be obtained from the authors upon request.
Algorithms
The following subsections shortly discuss the main characteristics of the three implemented algorithms. All algorithms have been used to predict the onoff times of appliances with a resolution of 1 h.
Histogram algorithm
Assuming that household activities follow a weekly pattern, one can build up a histogram of ontimes of an appliance for each weekday based on the training data (Chrysopoulos et al. 2014; Holub and Sikora 2013). The approach used in this work is shown in Eq. (1). It conditions relevant dayprofiles with a Gaussian weighting around the time of interest. In this manner we allow onevents in the past that are not precisely aligned with the time of interest to influence the prediction. Based on the preceding N days each subdivided into T time intervals, the probability that on day n at time t appliance l is running is calculated as
where x_{mlt} = 1 if appliance l was running during the interval τ on weekday m and x_{mlt} = 0 otherwise. w_{nm} = 1 if n = m, and w_{nm} = 0 otherwise. The variance σ is a model parameter that was set experimentally, see results section.
Pattern search algorithm
Whereas the histogrambased approaches assume the weekdays to be the governing pattern defining the weights w_{nm}, see Eq. (1), the approach by Barbato (Barbato et al. 2011) tries to identify these patterns. It does so by relying on the redundancy of variably sized daypatterns. To this end, one maps the N days preceding the day to be predicted, n, to a binary array of the form [δ_{n − N}, δ_{n − (N − 1)}, …, δ_{n − 1}] of length N with δ_{i} = 1 when the appliance was running on day i, and δ_{i} = 0 otherwise. The subarray S_{n}(i) is then defined as S_{n}(i) = [δ_{n − i}, δ_{n − (i − 1)}, …, δ_{n − 1}] for a given length 1 ≤ i < N/2. The occurrences of both the subpattern S_{n}(i) as well as [Sn(i), 1] = [δ_{n − i}, δ_{n − (i − 1)}, …, δ_{n − 1}, 1] is counted in the original array and the probability of a pattern of length i followed by a conjectured onday is calculated as
and correspondingly s_{n}(i, 0) for a conjectured offday (note that s_{n}(i, 1) + s_{n}(i, 0) = 1 by construction). Now i is increased until either s_{n}(i, 1) or s_{n}(i, 0) equals 1. In the latter case, a day without any appliance usage is predicted. Whereas in the former, the days following the occurrences of the pattern S_{n}(i) define the relevant days used for forecasting. They replace the days with identical weekday as used in the Histogram algorithm. It turns out that for the investigated data, patterns are not as obvious as in (Barbato et al. 2011), i.e. there is typically not an optimal pattern length i resulting in either s_{n}(i, 1) or s_{n}(i, 0) being 1. We therefore extended the original approach as can be seen in Eq. (3). Day n is predicted by the sum of the K most probable patterns weighted with the probabilities s_{n}(i, α).
where ∑_{i, α}goes over the K most relevant patterns. The Kronecker Delta δ_{α1} leads to a zero contribution of the patterns predicting a day with no appliance usage.
Bayesian inference algorithm
The third investigated method (Truong et al. 2013) uses Bayesian inference, which differs fundamentally from the previous approaches. It uses a MarkovChain MonteCarlo approach to sample the posteriori distribution of the model parameters. The key elements of the model are the latent daytypes k. They are used to create day profiles and to record correlations between the use of individual appliances. In summary, the probability p(x_{nlt}) of appliance l running at time t on day n is calculated as
where k goes over all K daytypes and p(k n) is the probability of day n being described by daytype k. One of the advantages of this approach is that it infers the parameters for each appliance l from the data of all appliances resulting with an effective training set of N · L data points, L being the total number of appliances.
Data and methods
Test data
Various datasets containing electrical consumption data of individual households are available (Murray et al. 2017). The three datasets employed in this investigation are GH9, collected by the authors, AMPds2 (Makonin et al. 2016) and REFIT, House 5 (Murray et al. 2017). They all cover at least two continuous years of data records from a singlefamily house, stem from Switzerland, Canada and UK respectively, and include submetered data for dishwasher, washing machine, and tumble dryer. In order to produce hourly on and offtimes off the appliances, the measurement data was preprocessed by imposing i) minimal on and offtimes i.e. removing noise spikes and preventing doublecounting due to intracyclic pauses, ii) as well as a minimal power levels. It was then downsampled to hourly intervals.
Performance metric
Binary classifiers can be assessed with a variety of performance metrics. We compare the predictive quality of the tested algorithms on the basis of the so called ReceiverOperatorCharacteristics (ROC) curves because of their independence from the relative weight of the groundtruth’s classes. The ROC method is well suited for a posteriori measure of the prediction quality but for an actual predictive algorithm a single working point along the curve (i.e. a single fixed threshold) must be chosen in advance. To average the ROC curves over individual samples each predicting the onoff behavior of an appliance during 1 week and estimate the resulting statistical variance, methods described in (Macskassy and Provost 2004) are employed. To allow for simple comparison with other experiments, the ROC curve is integrated, resulting in the area under curve AUC.
Results
Where not otherwise mentioned, results stem from an average over 90 samples, where each individual sample predicted onoff behavior of an appliance during 1 week based on the eight preceding weeks, hence covering in total roughly 2 years of data.
Histogram algorithm
The basic histogram method was tested on all three datasets with the model parameter σ (variance) varying between 0 and 2. Overall best performance with respect to AUC was achieved with σ = 1.3 which was used for all further experiments. The average performance improves by increasing the training window, i.e. increasing the individual trainsets, but saturates for lengths above about 3 months. As a tradeoff between prediction quality and a quickly increasing computational effort for the more elaborate algorithms, a training window of 8 weeks was chosen to ensure comparability of the results. Table 1 summarizes the results. The algorithm generally performs in a medium quality range with AUCvalues around 0.7. Differences in the AUC of different appliances and datasets are large but are not significant due to the large uncertainty as illustrated in Fig. 1.
Pattern search algorithm
Whereas the Histogram algorithm and the Bayesian Inference algorithm can ‘predict’ arbitrarily far into the future, the Pattern Search algorithm adapted from (Barbato et al. 2011) is only able to predict the day immediately following the training set. Thus for the latter, the window of training and validationset was not shifted by intervals of a week but day by day. Seven daypredictions were then summarized to a 1 weekprediction. Barbato’s approach has been modified to include the K most relevant patterns. Experimentally we found that K = 14 leads to satisfactory performance. As can be seen in Table 1, the AUC values are around 0.7 as for the Histogram algorithm but the standard deviation for the Pattern Search algorithm is mostly reduced.
Bayesian inference algorithm
In contrast to the results discussed so far, results from the Bayesian Inference algorithm originate from averaging not only over 90 samples i.e. validation weeks, but in addition each sample was obtained by averaging the prediction of ten independent Markov Chains. Tests showed a fast convergence of the individual Markov Chains independent of the initialization. With a burnin period of 500 steps, the individual prediction was calculated by averaging over 2000 Gibbs iterations. Results are summarized in Table 1.
Discussion and conclusions
The results summarized in Table 1 lead to the following observations: i) The overall performance of the three algorithms is essentially the same: averaging all appliances in all houses leads to the following values: 0.72 for Histogram and 0.73 for both Pattern Search and Bayesian Inference algorithms. ii) With increasing complexity of the algorithm not only the variances over 2 years decrease, but also the prediction quality across appliances and datasets becomes more similar. iii) An algorithm’s (relative) performance for a given appliance and dataset does not necessarily relate to another algorithm’s performance on the same dataset. That is, an algorithm performing particularly well on a given appliance of a given dataset does not necessarily imply other algorithms to perform similarly, i.e. the governing reason for the large differences does not seem to be the underlying data. iv) Similarly, no statement is possible about certain appliances performing markedly worse or better across all datasets for a specific algorithm, i.e. no particular algorithm is especially good at predicting a certain appliance. As discussed above the results imply that the mean predictive performance is not affected by the choice of the employed model. Because of the complexity of the implementation and computational considerations this would favor the Histogram algorithm over the two others. It performs at least 2–3 orders of magnitude faster than the algorithm based on Bayesian inference. For most realworld applications, it is, however, not the mean performance that counts most but a reliable performance on any given data. Here the reduced variance of the Bayesian algorithm with respect to the Histogram and Pattern Search approach speaks in favor of the former.
From our viewpoint, a limitation of the Bayesian algorithm in its current form is the fact that it strongly relies on weekly patterns despite its introduction of the latent daytypes. This could be addressed by making minor changes to include more data such as weather or schedules (Truong et al. 2013). An alternative could be to combine the Bayesian and the Pattern Search algorithms so that day n would also be predicted based upon the inferred daytypes of the immediately preceding days.
One aim of this study was to investigate if the onoff times of domestic appliances can be predicted solely based on electrical usage data. From our results, we tend to negate this hypothesis: On a coarsegrained timescale of 1 h, we achieved on average a mediocre prediction performance with a large variance. However, we believe that for domestic load optimization an improved performance and, in particular, a smaller variance of the prediction would be desirable if not necessary. Our choice of algorithms is far from exhaustive and one can think of various improvements of the examined algorithms. Nevertheless, from our point of view, the presented results based on 2 years of data from three different households reflect a general limit for the hourly predictability of an individual household’s electrical appliances. We conclude that taking solely electrical data of a single family into account, every stochastic approach must suffer from a lack of information, independently of its complexity.
Abbreviations
 AUC:

Area under curve
 ROC:

Receiver operating characteristic
References
Barbato A, Capone A, Rodolfi M, Tagliaferri D (2011) Forecasting the usage of household appliances through power meter sensors for demand management in the smart grid. In: 2011 IEEE International Conference on Smart Grid Communications (SmartGridComm). IEEE, Brussels, pp 404–409
Chrysopoulos A, Diou C, Symeonidis AL, Mitkas PA (2014) Bottomup modeling of smallscale energy consumers for effective demand response applications. Eng Appl Artif Intell 35:299–315
Holub O, Sikora M (2013) End user models for residential demand response. In: Innovative Smart Grid Technologies Europe (ISGT Europe) 4th IEEEPES, pp 1–4
Macskassy S, Provost F (2004) Confidence bands for ROC curves: methods and an empirical study. In: ROC Analysis in AI, First Int Workshop (ROCAI2004) Valencia Spain
Makonin S, Ellert B, Bajić IV, Popowich F (2016) Electricity, water, and natural gas consumption of a residential house in Canada from 2012 to 2014. Sci Data 3:160037
Murray D, Stankovic L, Stankovic V (2017) An electrical load measurements dataset of United Kingdom households from a twoyear longitudinal study. Sci Data 4:160122
Pflugradt ND (2016) Modellierung von Wasser und Energieverbräuchen in Haushalten. TU Chemnitz
Truong NC, McInerney J, TranThanh L, Costanza E, Ramchurn SD (2013) Forecasting multiappliance usage for smart home energy management. Proc TwentyThird Int Jt Conf Artif Intell Beijing China 4:3–9
Funding
This work has been financially supported though the Swiss Competence Centers for Energy Research – Future Energy Efficient Buildings and Districts. Publication costs for this article were sponsored by the Smart Energy Showcases  Digital Agenda for the Energy Transition (SINTEG) programme.
Availability of data and materials
The datasets analyzed during the current study are available from (AMPds2) Harvard Dataverse at https://doi.org/10.7910/DVN/FIE0S and (REFIT) University of Strathclyde’s PURE data repository at https://doi.org/10.15129/9ab14b0e19ac4279938f27f643078cec. The datasets GH9 is for the moment only available from the corresponding author on reasonable request.
About this supplement
This article has been published as part of Energy Informatics Volume 1 Supplement 1, 2018: Proceedings of the 7th DACH+ Conference on Energy Informatics. The full contents of the supplement are available online at https://energyinformatics.springeropen.com/articles/supplements/volume1supplement1.
Author information
Affiliations
Contributions
MG implemented the algorithms, and analyzed and interpreted their performance on the different datasets. The manuscript was written jointly by PH and MG and critically revised by both AR and AP. All authors read and approved the final manuscript.
Corresponding author
Correspondence to Patrick Huber.
Ethics declarations
Competing interests
The authors declare that they have no competing interests.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
About this article
Published
DOI
Keywords
 Load forecasting
 Shiftable loads
 Domestic appliance
 Experimental comparison