Go back to article: Visualising electricity demand: use and users of a 3D chart from the 1950s

The object in use

Our close study of the object, and the trends revealed, led us to speculate that its use was probably related to one or more of three important responsibilities of the electricity boards: forecasting, load-shifting and tariff-setting, and research and development. At this stage we used these lines of enquiry as keywords for interrogating the archive, to test our interpretations, and draw some conclusions about the significance of the object.

For forecasting?
Our first thought after examining the object and its data was, ‘Who would want to see the trends in electricity demand over time?’ Who were the CEGB ‘planners’? One possibility was those responsible for forecasting. Research in the Electricity Council archives at MSI gave an insight into this technical process. An article on long-term planning in the electricity supply industry noted that forecasts of demand were made for a number of specific purposes: for the operating programmes of existing plant, the provision of new plant, and the future strategy on the type and location of generating stations and transmission lines (Melling, 1964, p 3). As the design, construction and commissioning of a generating station took at least five years, forecasts in c.1954 would inform planning for the early 1960s.

In the late 1940s and 1950s there was much scrutiny of the methods of forecasting, not least because the industry had only recently been nationalised, and Area Boards were encouraged to use ‘common bases, principles and methods of estimation’ (BEA, 1950b, p 1). In addition, growth in demand during this period was consistently underestimated (Hancock, p 16), and the variation in winter weather could alter the maximum demand on the Grid system as a whole by at least a thousand megawatts (BEA, 1950b, p 16). As a consequence, the Load Estimation Working Party, reporting in October 1950, recommended that from the winter of 1950–51, Area Boards be urged to start: ‘the systematic collection of (i) daily loads at B.E.A. supply points and (ii) daily temperature readings (preferably at midday) at suitable positions in each Sub-Area’ (BEA, 1950b, p 17). (B.E.A. supply points were power stations or sub-stations at which the Area Board took a supply from the central authority, the British Electricity Authority).

Data was to be collected by the Area Board’s District Officers and entered into standardised forms, specimens of which were appended to the Working Party report. However, the forms were only concerned with the maximum demand on the Area Board, between 07:00 and 19:00 on working days, November to February (BEA, 1950b, Appendix B, Table 1). This indicates the window in which they expected to witness peak demand: during the working hours of a weekday, in winter months. The report noted that ‘Where practicable, separate records should be kept of morning, midday and evening peak loads’ (Ibid, Appendix A, p 5), but for the purposes of forecasting this level of detail was not required. Planning for engineering requirements (generating plant and transmission lines) appeared to need only data on the maximum demand.

At a more local level, Hughes (1983) describes the use of load curves by ‘load dispatchers’ – those tasked with controlling the supply from generating stations to meet demand. ‘Load forecasting’ required the retention of historical records, including factors directly affecting the magnitude, location and variation of the load, such as the weather (Hughes, 1983, p 367). In early twentieth-century Chicago, the load dispatcher was ‘not only a troubleshooter but something of a historian. At his disposal were load curves… By analysing these, the dispatcher anticipated loads resulting from the social customs and industrial routines of Chicago’s population’ (Hughes, 1983, p 214). Hannah (1979) also states how load dispatchers developed ‘a wealth of knowledge on the social life and working routines of the population they served’ (Hannah, 1979, p 125). Could this have been a use for the model in 1950s Manchester? As discussed above, and demonstrated in Figure 15, the trends in use from year to year across the period (1951–54) were not dissimilar, but spikes in demand were problematic. The electricity supply industry clearly had a need for data to help understand and manage load profiles and meet the fluctuating demands of the daily pattern, and this had a financial imperative as well as for resilience and long-term planning. This led us to our second speculation on the importance of the object for load-shifting and tariff-setting.

For load-shifting and tariff-setting?
Visualisations of the daily load curve, at thirty-minute intervals, gave information at a glance about the fluctuations in demand. Steep peaks in electricity usage in the mornings and evenings did not make for an efficient use of the supply system, and shifting the peak load to ‘flatten’ the curve made financial sense for electricity suppliers. Knowing exactly when those peaks occurred, and of what practices they were constituted, enabled the supply industry to think about ways to steer changes in the timing of demand, such as by promoting the use of night-storage heaters rather than on-peak electric fires (Carlsson-Hyslop and Pearson, 2012).

Another means of managing demand was the introduction of tariffs, enabling suppliers to set prices for more accurate forecasting, and to incentivise changes in electricity demand with cheaper off-peak tariffs. From 1948, one of the North Western Electricity Board’s primary tasks was the standardisation and simplification of tariffs for domestic, industrial, farm and commercial customers (Frost, 1993, p 68). In 1952, the Board introduced a new tariff for both domestic and industrial users, which replaced all the previous tariffs. During the decade 1951 to 1961 there were only two tariff increases, in 1955 and 1957, raising charges by fifteen per cent. But, as Hannah (1979) notes, the pricing policy of the industry had not been significantly modified until this decade (Hannah, 1979, p 203). An ability to see the times and seasons of peak demand at a glance would undoubtedly have benefitted those tasked with managing demand.

If the aim of the chart was to help intervene in patterns of societal synchronisation, shift the timing of what people do, and change the rhythm of the day, were these efforts successful? Recently scholars have suggested that dynamic pricing of energy has a limited effect on when people use energy (Torriti, 2015). This is because energy is used for social practices that are shaped by infrastructural and institutional boundaries and follow a fixed social rhythm (Shove and Walker, 2015). Against this, load models have lately been joined with time-use data surveys in thinking more carefully about what energy is used for – for example, what kinds of activities contribute to peak demand in the evenings. But, even today, as Torriti (2015) argues, the focus of demand-side management is persistently on tools that rely on changing energy demand through tariffs rather than drawing attention to why energy is used in the first place and which activities could be shifted from peak periods to off-peak periods.

Even if the object was intended to assist with load management, this does not fully explain the level of detail recorded on each card, which would have been hidden when assembled as a whole. Annotations such as the death of Queen Mary, and that meter readings had been verified (see Figures 12 and 7) convey date-specific information. These indicate the cards had a use before assembly, but whether those responsible for load-shifting and tariff-setting valued that level of data, and if so why, is difficult to discern.

For lobbying or demonstration?
The time invested in producing the model led us to consider a third possibility: that the object was related to lobbying or demonstration. Referring to scale models for performativity, Franck Cochoy describes how they are ‘less abstract than words, but more abstract than things…a bridge between language and practice’, and ‘embodied theories’ (Cochoy, 2010, p 308). There is the simple fact that the model might have been ‘putting an idea to work’: ‘showing’ being better than ‘just talking about it’.

As noted above, the model dates to the end of the analogue era. A recollection by R G Hancock, Head of the Intelligence Section of the Electricity Council, notes that computers were being used on the CEGB’s data sets from around the early 1960s (Hancock, p 16A), so perhaps the construction of this model was a means of lobbying for such resources? Perhaps it was to prove that creating this data visualisation with manual methods took this many hours/days? An organisational change in 1954, from the British Electricity Authority to the Central Electricity Authority could also have been a driver for lobbying and/or research, e.g. to show the Sub-Area’s demand profile to a new audience.

This is supported by the evidence that the model was highly esteemed, and of its possible role as a showpiece. The green felt pads on the underside suggest it was destined for a boardroom table. There was also the expense of the colour-printed cards, and the time invested in transferring data onto the cards and assembling the whole. It was also probably ‘on display’ for some time. The degree of fading where cards were exposed to the light is broadly consistent across the whole (see Figure 17). An ‘interpretation panel’ stored with the object (see Figure 18) suggests that for at least some of its lifetime it was used for instructing non-specialist audiences. Whether the object performed these roles at the time, or subsequently, is difficult to distinguish.

Figure 17

Colour photograph of an individual chart card from a 1950s three dimensional chart showing electricity demand over time

Significant fading of the card is noticeable in the time period 12:00–17:00

Figure 18

Colour photograph of an individual chart card from a 1950s three dimensional chart showing electricity demand over time

A laminated interpretation card stored with the object, with two cardboard stands fixed to the reverse side

Component DOI: http://dx.doi.org/10.15180/180905/004