Modular, relocatable solar farms: demonstrating a new asset type to increase residual value and enable a new wave of solar users
To date, the deployment of solar farms may be characterised by long-term power purchase agreements (PPAs) with credit-worthy offtakers such as utilities and government entities.
Authors: Dr Will Rayward-Smith; Mr Lewis Cowper; Mr Paul Ward; Prof Andrew Harris; Mr Michael Read
In locations, such as remote islands, where labour resource is sparse or expensive, how can we deliver solar affordably? In sectors, such as mining, where clients want to contract power for short terms, how can we make solar financially attractive? How can we lower the project-specific engineering and development costs to make solar more cost competitive? Modularity provides answers to these questions, by lowering on-site activity, enabling relocation and facilitating an off-the-shelf scalable design. The idea of modular solar was conceived in late 2013. Modular solar involves integrating multiple solar panels together with simple wiring looms into identical, container-sized modules off-site; stacking these modules for efficient transport to site; and then safely and efficiently assembling on-site. With the support of the Australian Renewable Energy Agency, the multi-panel module was developed in 2014 and piloted at a remote construction village in 2015. This article introduces the modular philosophy to solar farms, shares lessons learned from implementation and provides an overview of performance of the pilot plant. With the success of the pilot, and confidence in both the value proposition and addressable market, the technology and business model have undergone further product development and has now entered commercialisation, with the launch of SunSHIFT TM Pty Ltd.
To date, the deployment of solar farms may be characterised by long-term power purchase agreements (PPAs) with credit-worthy offtakers such as utilities and government entities. In this context the value at the end of the PPA, the residual value, is largely inconsequential in the eyes of financiers. This lack of requirement around residual value has led to a product that is stick-built and immobile. Such a solution does not meet the needs of all segments of the market and is limiting the uptake of solar.
The off-grid sector is characterised by temporary or less credit-worthy entities that are unable to commit to long-term PPAs. Short mine-life mines are unable to commit to long-term PPAs, even if it is likely that the mine-life will be extended. Even long mine-life mines prefer shorter term contracts for their power to maintain operating flexibility and keep assets off their balance sheets. Remote utilities are nervous about committing to long-term PPAs in case their user demand changes dramatically with increased uptake of behind-the-metre solar or population movement. In such scenarios, capturing greater residual value is critical for the securement of finance.
Modularisation of a solar farm, incorporating tens or hundreds of components into single modules off-site, lowers costs associated with assembly and disassembly to facilitate relocation, thereby enabling greater residual value to be considered in financing models. Furthermore, a modular solution is inherently scalable, reducing project-specific development and engineering costs to meet client-specific requirements. A modular approach can be demonstrated to be capex comparable to stick-built solar farms, particularly in remote areas where mobilisation and labour costs associated with on-site activity are high. To summarise the benefits of this approach, modular solar is compared with conventional, stick-built solar in Table 1 .
CONVENTIONAL, STICK-BUILT SOLAR COMPARED TO MODULAR SOLAR
|On-site activity||high man-hours||low man-hours|
|Speed of deployment||long programme||short programme|
|Relocation||complex/costly to relocate||easy/low-cost to relocate|
|Residual value||zero/low residual value||high residual value|
|Capex recoupment period||offtake duration||asset life|
|Viable PPA duration||only long PPA terms possible||short PPA terms possible|
Table 1: Conventional, stick-built solar compared to modular solar
In this article we demonstrate the deployment of a large-scale relocatable solar farm and how it was successfully transported, assembled, integrated with existing diesel generators and commissioned in less than one week. Performance of the solar farm is also presented along with a brief discussion on future development of the technology.
The design philosophy has been one of Design for Manufacture and Assembly (DfMA), seeking to maximise the benefits of off-site manufacture and reduce on-site activity to a safe and efficient assembly procedure. The modular nature of the product also enables low-cost development and engineering design to meet customer requirements over a broad range of system sizes.
To streamline the product development phase, each type of SunSHIFT module has its own digital model that was meticulously maintained during design. Changes to any model were reviewed by all engineering disciplines to ensure that overall improvements were made. For example, the relationship between solar panel tilt angle and subsequent wind loading must be considered. Such interdependencies can be explored efficiently within the digital model.
The total power plant is comprised of five module types:
- Solar modules
The solar modules are identical, each consisting of frame structure, electrical wiring loom and multiple solar PV panels (Fig 1). The solar modules are rated to withstand cyclonic wind loadings with the number of ground restraints per solar module determined by the wind region of the deployment location.
Fig1: Solar modules
Solar modules are grouped with a string inverter to form an Array. String inverters (Fig 2) are used rather than central inverters for their superior reliability, redundancy, energy yield, maintenance and warranty. Pre-terminated fly-leads are used to electrically connect the solar modules and string inverter within the Array. There are different types of fly-leads to provide the necessary parallel and series distribution within the Array.
Fig 2: String inverter
In megawatt-scale deployments, Arrays are grouped into Blocks. Each Block has a transformer that then connects to a high voltage (HV) ring main. Multiple Blocks may connect to the HV ring main to form a Farm.
- Gas/diesel generator
The majority of deployment opportunities are brownfield with existing diesel/gas power generators in place. As such, the relocatable solar farm has been designed to augment both new and existing generators of any size and configuration.
- Fuel tank
The Integrator houses the electrical switchboard and control system (Fig. 3 ). The control system monitors the generators and site loads to maximise solar power output whilst maintaining efficient operation of the gas/diesel generator fleet. Additionally, the control system ensures that sufficient spinning reserve, either from the generators or energy storage, is available to manage demand spikes and cloud cover events.
- Energy store
Fig 3: Integrator
Higher levels of solar PV penetration may be enabled with the addition of a containerised energy stores to the system. Additionally, energy storage can increase generator fuel efficiency by supplying spinning reserve capacity instead of the generator fleet.
A pilot-scale relocatable solar farm was manufactured in late 2014 and deployed at a 550-bed construction worker accommodation village in remote Queensland, Australia in February 2015. The pilot plant was operated for a six-month trial period before demonstration of the pack-up phase in August 2015. The plant provided all electrical power to the village and its vital statistics are summarised in Table 2.
|Optimised configuration||144 × 1 kW DC solar modules (grouped into 6 × 20 kW AC arrays)|
|3 × 400 kW AC and 1 × 300 kW AC diesel generators|
|1× diesel tank|
|0× energy store|
|Solar panels||SunPower E20/327|
|Height of modules off-ground||due to the short deployment duration (i.e. low vegetation risk) and existing grading/sloping of the site (i.e. low flood risk), the modules were able to be placed on the ground|
|Control system||ABB Microgrid Plus|
|Wind region (AS/NZS 1170.2)||A|
|Soil condition||ground is generally underlain by up to 200 mm of high plasticity clay fill of hard consistency, overlying up to 1.2 m of natural high plasticity clay of firm to stiff consistency. The natural clay is overlying extremely weathered sandstone of very low to low strength|
|Solar module screw-pile density||placed between every third solar module (due to low wind loading and favourable soil conditions)|
|Forecast solar energy export||180 MWh p.a.|
|Forecast diesel saving||45,000 L p.a.|
|Forecast CO 2 emission reduction||120 TCO 2e p.a.|
|Commercial arrangement||6-month hire|
Table 2: Vital statistics of the pilot plant
The pilot plant included multiple steelwork items that were manufactured off-site in Newcastle, Australia (Fig 4). Laing O'Rourke engineered the inverter modules and Integrator to have the same steel frame in order to reduce engineering and design costs and increase economies of scale during manufacture. The steelwork included seven such frames (six inverter modules and one integrator) and 144 3-panel solar modules.
Fig 4: Fabrication of solar modules
All equipment was transported from the fabrication facility in Newcastle, Australia to site in 20 ft open-top shipping containers. This mode of transport was selected to minimise the risk of damage to equipment during transit whilst conforming to all road transport dimension limits.
Assembly of the system consists of two major packages – screw-pile installation and solar module placement. Minor packages include inverter and Integrator placement and electrical connection of all equipment, including diesel generators.
For the pilot system, 66 screw-piles were installed using a 13 T excavator fitted with a torque head (Fig 5).
Fig 5: Installation of a screw-pile
Placement of the modules (Fig 6) was achieved in three days with a crane truck and three-man crew (operator, dogman and labourer).
Fig 6: Solar module placement
The remaining works – placement of inverters (Fig. 7 ), placement of the Integrator and electrical connection of all equipment – were completed in parallel with the module placement package.
Fig 7: String inverter placement
The 120 kW AC system was installed and commissioned in under one week (Fig 8). Time-lapse footage of the assembly is available to the public via the Australian Renewable Energy Agency (ARENA) YouTube channel – https://youtu.be/-ZOR1trhJD8 (accessed February 2016).
Fig 8: Aerial photograph of the accommodation village with the completed solar farm (located in the near right corner of the site)
Assessment of the performance of the pilot system is divided into four categories.
In late 2014, PV simulator software was used to forecast the output of the solar farm. The simulated performance and the actual performance are presented in Fig 9. There is strong agreement between the simulated and actual performance showing that the SunSHIFT system performed as expected.
Fig 9: Cumulative forecast and actual energy output of the solar farm
Effort was made to maximise the solar energy contribution of a system with low solar power penetration by moving time-shiftable loads to be in-line with the solar output. For example, hot water heating is a load easy to shift to the daytime.
Integration and control
Although maximising the energy output of the solar PV is favourable, it is important that the system protects the diesel generators by running them at or above their minimum loadings. As such, if an online diesel generator is running below its minimum load and there is available solar power, the system is designed to curtail solar power and run the generator at its minimum load. Although an energy store was not included in the 120 kW AC pilot system, if energy storage is present then excess solar energy may be stored instead of curtailed.
We now look at a particular example to demonstrate that the system is curtailing solar correctly. Our example of solar curtailment occurred at 14:12:20 on 1/6/2015 as is described with reference to Fig 10 as follows:
- At approximately 14:12:21 there is a 20 kW reduction in the total demand.
- During this event, the online diesel generator (Gen 3) drops below its target minimum loading. The Integrator responds by curtailing the solar PV output and increasing the loading of the diesel generator.
- At 14:12:24, just 3 s after the change in demand was identified, the diesel generator is back above its target minimum loading.
Fig 10: Example curtailment of solar output to meet the minimum load requirements of the online diesel generator
This example demonstrates that the control system of the plant is curtailing solar correctly in order to protect the diesel generators and effectively manage the integration of diesel and solar PV power generation.
Reliability of supply
Shading from passing clouds, referred to as cloud events, is a common factor impacting the reliability of supply from solar PV. It is critical that voltage and frequency remain stable during such cloud events. The system ensures this by maintaining sufficient spinning reserve in the diesel generator fleet at all times to cover any sudden loss of solar output and spike in demand.
The system was originally configured to provide static spinning-reserve of 80 kW for step loads from the village plus a dynamic spinning-reserve equal to 100% of the real-time PV output for cloud events. During commissioning on-site, however, it was observed that even on a day with complete cloud cover the PV still produces approximately 20% of its rated output. As a result the dynamic spinning-reserve requirement was reduced to 80% of the real-time PV output.
We now look at a particular example to demonstrate that the system is maintaining spinning reserve correctly. Our example cloud event occurred at 13:06:30 on 5/4/2015 and is described with reference to Fig 11 as follows:
- From 13:00 to 13:06, there are several cloud events that cause the solar output to decrease. During these cloud events, the online generator (Gen 3) increases its power output to meet the demand.
- A cloud event reducing solar PV power output by 90 kW begins at 13:07:05. Over the next 20 s, demand also increases by 100 kW. Gen 3 responds by increasing its power output from 202 kW to 388 kW over 23 s.
- Although the system is stable, there is now insufficient spinning reserve in the system. In response, the Integrator instructs a second generator (Gen 2) to come online.
- At 13:07:40, Gen 2 comes online and as it is of equal size to Gen 3, they are configured to share the load equally.
- The cloud event ends at approximately 13:11:00 and solar PV power output returns to an expected level. Whilst a single generator is able to cover remaining demand plus the spinning reserve requirement, Gen 2 has not satisfied its minimum on-status duration and so both generators remain online.
- At 13:20:52, Gen 2 meets its minimum on-status duration and is shut down. Simultaneously, Gen 3 is ramped back up to meet demand and spinning reserve requirements.
Fig 11: Performance of the solar-diesel hybrid system during a severe cloud event
The hybrid system reduces diesel fuel usage in two ways:
- The output of the solar PV displaces diesel fuel.
- The Integrator offers fleet management services, optimising the use of the diesel generators.
We can look at the fuel efficiency of the pilot plant to determine its effectiveness. Relative to an adjacent camp where the diesel generators had a very poor fuel efficiency of 0.355 L/kWh, the pilot plant reduced diesel fuel consumption by 74,700 L in the 81 day period from 21/02/2015 to 11/05/2015 (inclusive), equivalent to reducing emissions by 166 TCO 2e (Fig. 12 ).
Fig 12: Fuel consumption rate (L/kWh electricity exported to the mini-grid) of the SunSHIFT pilot plant. The whiskers of the box are the minimum and maximum (excluding outliers). The ends of the orange box are the first and third quartiles and the band inside is the median. The blue marker represents the mean fuel efficiency over the time period
During this time period:
- The SunSHIFT plant exported 588,056 kWh of electricity to the mini-grid, of which 55,933 kWh came from the solar farm – i.e. a solar energy contribution of 9.5%.
- The inverters consumed 57.7 kWh of electricity for their operation, equivalent to just 0.1% of the solar electricity exported.
- The Integrator monitored the diesel fuel consumption rate of each of the diesel generators and metred a total of 134,138 L of diesel fuel consumed.
- The fuel consumption rate for the SunSHIFT pilot plant was therefore 0.228 L/kWh.
To measure the acceptance of large-scale solar on a remote work site, Laing O'Rourke conducted a survey with 58 blue-collar workers and 26 white-collar workers within the accommodation village that is powered by the pilot plant. Participants were asked (anonymously) to score five priorities from low to high importance. The average scores of the priorities is shown in Fig. 13 . The results reveal the importance that those individuals place on reducing carbon emissions above other objectives.
Fig 13: Average scores of five priority areas from a survey conducted with a sample of the accomodation village community
Participants were also presented with two statements related to large-scale solar on a remote worksite and asked to select responses that most accurately reflect their own feelings. The results of this section of the survey are presented in Tables 3 and 4 and show strong support for large-scale solar on remote worksites.
|RESPONSE||COUNT||PERCENTAGE OF TOTAL, >#/TH###|
|‘Good to work for a company trying new tech on my project’||70||83.3|
|‘Negatively because it takes up a lot of space/is not linked to the project’||1||1.2|
Table 3: Responses to the statement ‘The re-deployable solar farm at the Combabula site is the first of its kind in the world. Does this make you think’
|RESPONSE||COUNT||PERCENTAGE OF TOTAL >#/th###|
|‘I don't know’||19||22.6|
Table 4: Responses to the statement ‘Solar-diesel power plants have the potential to positively change how we generate and use power in remote Australia ’
Energy and carbon payback
The mass of hot-dip galvanised steel in a module frame is approximately 335 kg. Combining a diesel fuel efficiency rate of 0.25 L/kWh with the World Steel Association Life Cycle Assessment Methodology (2011, ISBN 978-2-930069-66-1) values for the energy intensity and carbon intensity of hot-dip galvanised steel, 27.5 MJ/kg-steel and 2.5 kgCO 2/kg-steel, respectively, an energy payback of 7 months and carbon payback of 10 months was calculated for a module frame.
This demonstration has shown that modular, relocatable solar is technically feasible. Further development of this technology is chiefly dependent on its commercial viability.
Engagement with the mining sector, remote communities and emerging economies has led us to believe that there is a substantial addressable market. Off-grid miners in particular are interested in solar but conventional offerings do not meet align with their preference to contract power short-term to maintain operational flexibility.
Our value proposition – that is, offering the environmental and financial benefits of solar electricity generation with the speed of deployment, flexibility and reliability of the industry-trusted diesel generator – has resonated with each of the potential offtakers that we have engaged with. The addressable market recognises the opportunity to divert expenditure on diesel fuel towards investment in sustainable power generation assets such as SunSHIFT.
It is difficult and expensive to relocate a conventional, stick-built solar farm. This means that the capital cost must be recouped within the original offtake term. For short-term offtake agreements, this results in a high cost per unit of energy which is not viable for the offtaker.
It is easy and low-cost to relocate a modular solar farm. This means that the capital cost can be recovered over the asset life rather than having to be recovered over the original offtake term. Therefore a modular solar farm can be offered on short-term PPAs at a unit rate that is viable for the offtaker. Modularity decouples asset life from offtake duration.
The concept of a modular, relocatable solar farm has been successfully piloted, demonstrating its technical feasibility. Installation, operation and removal of a 120 kW AC pilot-scale solar farm was conducted on an off-grid accommodation village in remote Queensland, Australia. The system performed as expected, with no faults arising during the six-month trial operation period.
Engagement with miners and remote communities in the off-grid electricity market has encouraged further development and commercialisation of the modular, relocatable solar farm technology. Laing O'Rourke has launched SunSHIFT TM Pty Ltd.
The authors wish to thank all of the employees within Laing O'Rourke that have added their expertise to the success of SunSHIFT.
On behalf of Laing O'Rourke, the authors acknowledge ARENA for its vital funding contributions and ongoing supportive collaboration that has enabled this technology to develop.
On behalf of Laing O'Rourke, the authors acknowledge ABB and SunPower, the major technology partners for this demonstration.