Results of data collection for Pangboche
This stage is in accordance with second stage of the proposed methodology: Data Collection. This stage follows on from Project Definition and is followed by Energy System Design
As the project aims to design a system that will meet Pangboche’s electrical and thermal needs, data was required for inputs into modelling software. This would allow the systems to be viable for both the community and its needs. This section details the methods and results from data collection for Pangboche, according to Stage 2 of the proposed methodology.
Due to the inability of conducting a site visit, data was collected via literature and online resources in addition to some of the data provided by the village contact. Following the methodology, local sources were looked at first before moving on to national databases.
Below are the results of the data collection for designing renewable energy systems for Pangboche.
Run of river hydro [2]
The following data was identified as necessary to assess hydropower potential:
Pico/micro hydro (<100kW) is an established renewable technology in Nepal, having been introduced formally in the 1960s [1]. These were initially no more than small and basic water mills for agricultural use, later progressing to generators for electrification. For run-of-the-river pico/micro hydro systems, a portion of a river's water is diverted to a channel, pipeline, or pressurized pipeline (penstock) that delivers it to a turbine or waterwheel. The water rotates the wheel or turbine, which spins a shaft. The motion of the shaft can be used for mechanical processes, such as pumping water, or it can be used to generate electricity [2].
For a river to be feasible for a hydro system, it requires good head and good flow. Higher head is better, as less water is required to produce a given amount of power, therefore smaller and less expensive equipment can be used. For an effective system, head should be greater than 3m [3]. To assess this, data can be collected from meteorology or irrigation departments run by the national government or measured by hand if access to the area is possible. The main river by Pangboche is the Imja Khola which runs along the valley floor.
Google Earth Pro was used to estimate the head of the Imja Khola over a 1km distance from Pangboche and it was discovered to be very steep with an average rise of 0.08m height/m distance.
Flow was a challenging parameter to find, as the Khumbu valley is so remote there is little data on the flow of its rivers. In this situation, the most accurate method would be to self-measure the flow: it can be measured easily by the time it takes to fill a 1 m3 bucket or by the time it takes an object to float a certain distance. However, in the case of Pangboche, there was no physical access to measure the rivers. A local dataset was found for the total monthly flowrate of the Dudh Kosi river, which is further upstream from the Imja Khola [4]. The dataset could be corrected to show the average flow of the Imja Khola based on river depth and gradient. The steeper the gradient and the greater the river depth, the faster the river flows. The graph below shows the average monthly flowrate of the Imja Khola:
Rainfall data is useful for understanding how consistent the flow is likely to be. The Italian run EvK2-CNR Share Project, monitors high altitude metrological data and has several Automatic Weather Stations (AWS) located along the Khumbu Valley [5]. As shown in the chart below, Pangboche has good rainfall in the summer monsoon months but has very little rainfall over winter which could be a risk to a hydro power scheme.
The data collected in this stage was then used to determine the technical potential of hydropower during the system design stage.
The following climate data is identified as necessary to assess the technologies of PV and wind turbines as well as thermal considerations:
It was challenging to find a detailed resource assessment for the solar and wind energy available at Pangboche as the climate there has not been studied extensively. However, there was some data available on the climate of the Khumbu Valley (in which Pangboche lies) recorded by the NREL’s Solar and Wind Energy Resource Assessment project (SWERA) [6]. There was also some data from World Weather Online [7] that was used to define the climate of Pangboche. Local weather stations run by Italian research project EvK2-CNR were also used.
As can be seen, the solar resource in Pangboche is very strong throughout the year, however there is a great drop in direct normal irradiance (DNI) during the middle of year. This was expected as during this part of the year, Nepal goes through its monsoon season which would result in the sky being very cloudy. This also results in an increase of diffusive light during this time of year.
As Pangboche lies in a deep valley, the topography was considered to have a large impact on the solar resource of the area. To quantify the affect the landscape had on the solar irradiance throughout the day, the software Meteonorm was used to generate an image of the horizon from the perspective of the village and the solar path throughout the year.
An alternative to this software would be PVsyst, which could be used in conjunction with HeyWhatsThat which would produce the horizon profile. As can be seen from the screenshots provided, both software produced similar horizon profiles.
Meteonorm software
Horizon profile generated using HeyWhatsThat
Whilst this method of processing was explored for Pangboche, unfortunately time constraints did not allow the full utilisation of the results obtained by the software.
The temperature and relative humidity of the location was required to help define the current living conditions for thermal demand optimisation.
Pangboche gets very cold during the winter months, the average temperature being well below freezing for almost half of the year. Relative humidity also follows the same trend, which contradicts what one would expect as when temperature increases, relative humidity should decrease. This, however, makes sense when considering Pangboche will be experiencing a monsoon during the periods when the temperature increases.
As wind turbines were a technology that was considered, the average wind speed in Pangboche was also examined. So far, wind is a highly unexploited resource in the Khumbu region, as wind field evaluation is very challenging due to the highly complex morphology in the area. It was found that wind speeds were higher during the winters and lower during the summer monsoon. This does not show a good fit to the community’s needs as the population at Pangboche (and thus, the electrical demand) decreases during the winter as people migrate to Kathmandu.
This data would have been enough to start modelling the electrical system (Homer only requires monthly averages), but it was not possible to use this for building simulation. The software used (Esp-r) required an hourly data input for each parameter. To solve this, hourly climate data from the closest location was used and transformed to fit Pangboche’s climate. Hourly recorded data from Kathmandu was available from EnergyPlus via the National Renewable Energy Laboratory’s extensive SWERA database. To ensure the solar resource data form Kathmandu would be similar to Pangboche, the landscape of the village was mapped using Google Earth Pro and used as the input for Meteonorm. Meteornorm then used the landscape to calculate the hours and the amount of solar irradiance the area recieves throughout the day.
As climate data can vary greatly between regions of different elevation and topography, data collected from nearby areas may need to be transformed so as it better represents the climate of village. The closest source for hourly data was Kathmandu (the capital of Nepal). There is a large difference in elevation (and thus, climate) between the two locations, so the Kathmandu location had to be transformed to fit to Pangboche’s climate.
An example will now be shown using relative humidity. The relative humidity data collected from Kathmandu shows that relative humidity is higher in Kathmandu than Pangboche during winter, as can be seen in the graph below:
Initially, a trend-line of the Kathmandu data was plotted and using its equation, the data points of the trend-line for each hour were produced. To be able to use the Pangboche data, it had to be converted into hourly data. It was assumed that the average monthly relative humidity was reached by the end of each month to simplify calculations. This produced the following trend-line equation:
This equation was then used to produce hourly data points for the Pangboche average. This graph alone could have been used for modelling, but it would not be representative of Pangboche’s true climate. Thus, the stochastic nature of the Kathmandu data was used to make the results more realistic.
To achieve this, the difference between the Kathmandu data points and its trend-line were added onto the corresponding Pangboche trend-line data point. The result of this is that the Pangboche hourly data points will now appear stochastic in nature.
As can be seen from the graph above, the data now follows the general trend of the Pangboche average climate data. Clearly, this is not a perfect solution as the relative humidity in Pangboche may not vary as much as it does in Kathmandu. It can also be seen that the trend-line of the hourly data does stray a little from the average climate data. This is due to the assumption that the relative humidity reaches the Pangboche average by the end of the month. This difference is negligible as the data is only required to produce a realistic result from modelling. There was a little housekeeping that was done to produce the graph above.
Whilst the following will not apply to most parameters as there would not be a hard limit on possible results, it is still important to consider for parameters such as relative humidity. Relative humidity, as a percentage, has a lower limit of 0 and upper limit of 100. However, when implementing this method, it was found that as most of the Kathmandu data points were already close to 100%. Thus, when the points during the winter were brought down to Pangboche levels, it produced a flat looking graph. To fix this, the random number generator function in excel was used to produce a random number between -5 and +5, which was then added onto the hourly Pangboche data points. This random modifier helped the data appear more stochastic in nature. Furthermore, it was found that many points throughout the years were going beyond the appropriate limits for relative humidity. Thus, the minimum and maximum functions in excel were used to limit any values for relative humidity to be between 0 and 100. This method required the assumption that the variance of the parameter does not change between the two locations. This is not always the case and will depend on a case by case scenario. For example, the temperature throughout the day may only vary by around 15 degrees in Glasgow but it would vary much more in the Sahara Desert. To get the best estimates, the two locations should have a similar climate.
As the technologies identified to meet the cooking and heating needs of Pangboche were cook stoves and biofuels, the next step was to determine what data was required to assess the potential supply available for these technologies. In order to identify the fuel sources available to Pangboche, a meeting was conducted with a community contact to assess the sources used currently for cooking and heating. The information gathered from the community contact was that biomass sources such as animal dung (mainly Yak) and fuelwood were currently used for cooking by being burned on an open fire using traditional cook stoves. However it was also highlighted that due to Pangboche being situated in Sagarmantha National Park fuel wood collection was restricted to 10 days a year, during these 10 days only two bundles of fuelwood were allowed to be collected per day. Due to the shortage of fuel wood the community had to buy kerosene from Kathmandu for use as cooking fuel.
Other information which was collected from the local contact includes the number of residents in Pangboche, number of households and the number of people per household. With this information the next step was to determine the possible supply of energy from biomass which would be available to the community of Pangboche.
With the information collected from the community contact, further data was required on specific data such as livestock numbers for supply of animal dung, fuelwood available per household and other sources of waste which could be used for biofuels and cook stoves. In order to gather this specific data, the best method would be to carry out a market assessment on the community or a questionnaire per household which would gather more accurate information on specific community data. However as not all information required could be gathered from the community contact a literature review was carried out to acquire the missing information required for biomass supply.
In order to assess the animal dung available in Pangboche the following information was required:
The livestock mainly owned by Pangboche residents was mentioned by the community contact as being Yak as due to the high altitude of Pangboche few other mammals can live at such heights [1]. This was confirmed through a literature review which identified that the majority of livestock within Pangboche were Yak [2]. The next step involved identifying the number of livestock owned by Pangboche which again would be best assessed through a questionnaire of the community. However, as there was difficulty in finding extra information through the community contact, literature was reviewed to find the number of livestock owned by Pangboche. Through literature there was difficulty in finding current livestock numbers owned by Pangboche with the only relevant data found being livestock numbers in Pangboche in 2008 [2]. The graph found through literature that shows the livestock types and numbers from 1957-2008 in Pangboche has been included below [2]:
Using this graph the number of livestock in Pangboche in 2017 was extrapolated from following the trends shown in previous years. The graph above shows a growth of 207 livestock from 1984 to 2008, this is an average of 8 livestock per year. Assuming the same rate of growth from 2008 to 2017, this equals to 72 extra livestock per year which totals to 597 livestock in 2017.
With the number of livestock in Pangboche identified, the next step was to determine the number of households to calculate the number of livestock per household. Google earth was used at this stage to identify the number of households in Pangboche. In order to gain more accurate results ArcGIS a satellite software was used to identify extra households not determined by google earth, this came to 116 households. In order to determine the total energy supply from dung, the dung production per animal and energy content of dung were determined. This was done through literature reviews where the dung production per animal was found to be 10 kg/day [3] with an energy content of 10 MJ/kg [4]. Assuming the total dung recoverable was 70% to account for rainy days during monsoon season where dung would not be collectible and to account for losses, the total energy content from recoverable came to 41,790 MJ/kg/day. Therefore the supply of energy from dung per household came to 360 MJ/day, the calculations performed above are summarised in the table below:
In order to determine the energy supply from fuelwood for each household the data collected through the local Pangboche contact was used as a guide to determine wood collected per year. As the contact had mentioned that the Pangboche community are only allowed to collect wood 10 days per year, a review of literature showed that an estimated 30 kg of fuelwood is allowed to be collected per day per person [1]. With an energy content of 15.5 MJ/kg for fuel wood, the total supply of energy from fuel wood per household came to 25.5 MJ/day. This is summarised in the table below:
Although Pangboche has a large number of livestock allowing a large supply of energy through dung, other waste sources were assessed in order to increase the possible energy supply. In order to manage household waste in Pangboche the amount of organic food waste per household was determined through a review of literature. It was found that the mean household waste for a family of five in Nepal was 1000g/day, of which 71% was organic waste [5]. Using an average energy value of 9 MJ/kg for waste, the total energy supply calculated from organic waste per household was calculated to be 6.4 MJ/day [6].
Using the energy supply calculated for each biomass source, the total biomass energy supply mix per household came to 392 MJ/day, this is summarised in the table below:
Housing stock data is required to create a suitably detailed representation of a typical dwelling. This model can then be simulated with fabric improvements in place to estimate their impact upon the thermal performance of the building.
The modelling software used was ESP-r. This required the following data:
As no site visit could be conducted housing stock review and data gathering was conducted via literature review, discussions with the contacts, and some other methods which are discussed below. Housing stock review was performed via literature review and Google Street View (provided by a walking camera).
The basis of the model is its dimensions. Position, size and numbers of windows and doors are also required. Slope of roof and overhangs or shading are also included.
A typical house is shown in the figure below and consists of two storeys. The ground floor is used for keeping livestock at night (their body warmth provides heat for upper levels) and the upper storey is occupied by the family. A large portion of houses are semi-detached and built into hills (i.e. there is both a lower entrance and upper entrance on different sides of the house) however the model is neither.
The house was chosen via Google Street View and located through satellite imagery on both Google Maps and ArcGIS. House footprint dimensions were calculated using the imagery with a tool used to extract numerical data from graphs (specifically WebPlotDigitizer was used). Dimensions were found to be approximately 12m in length by 5m width. The storey height was found to be approximately 2.5m by comparing dimensions and roof slope of 30°. Eaves significantly overhang as a solar-shading measure.[1][2][3]
Constructions determined for the houses are listed below. Houses were found to be constructed of stone walls with internal wood paneling, external cement-screed, and no insulation materials throughout. The ground floor base is compacted earth. Floors and ceilings are timber as are doors and window frames. Windows are single glazed. Roofing is corrugated iron and timber. Loft insulation is not commonplace in the village with typically less than 50mm of glass insulation used.
Walls: 10mm dry white render, 450mm limestone, 20mm air gap, 20mm pine paneling
Ceiling: 20mm softwood, 50mm air gap, 20mm softwood
Roof: 3mm corrugated iron, 2mm EPDM membrane, 30mm pine
To model thermal gains from occupancy typical household use was generated from the modeler’s judgement. The district’s (Solhukumbu) average number of occupants per household of 4.5 was assumed [4].
It was found that households are used for a variety of activities including moderate work. Additionally, livestock (2-3 yak) often occupy the ground floor at nighttime so contributing heat to the household.
Thermal comfort temperatures were estimated from the literature to be as per the below table. Thermal comfort is noted to be very different to European standards. This is likely due to minor physiological differences of the Sherpa people and cultural adaptations, this is explored more in the following paper. [5] [6]
Loft and walls were identified as key areas for insulation. According to the contact, both are typically poorly insulated if at all (less than 50mm of glass wool). Ceiling insulation depths of 100mm, 200mm and 300mm are considered. Wall depths of 200mm only are considered to reduce simulation time.
As a site visit was not possible this data was not collected. During modelling, thermal bridges were later assumed as per the building modeling software. Air changes were assumed to be relatively high based upon the generally poor build quality of the houses. It was noted from the discussions with the contacts that cementing between stonework is done minimally due to the high import cost of the material. Additionally, window and door frames are rudimentary so further contributing to air leakages.
The next step of the case study in Pangboche is Energy System Design