This is a project that i just finished that is relevant.
Associated spreadsheet unfortunately not live but at the optimized state for NPV.
The scope of this project is a one house system, through other houses could be easily added, and only looks at electricity production. Only current off the shelf technologies are looked at, no prototypes or possible systems are included.
This project is looking at the optimisation of a microgrid that is only powered by diesel fuel or solar PVs.
Batteries are used as an energy storage medium and hot water is produced by the diesel generator but this
hasn’t been modelled.
Fixed into the system is the battery (specifically chosen) bank, load, generator model & type along with
Solar panel model. What can be changed is the number of solar panels.
Left out are other renewables; wind is ignored because it is inefficient at a small scale while Melbourne’s
wind resources aren’t sufficient to make it worthwhile and Biomass isn’t suitable to a suburban or urban
Optimization is aimed at an improvement over connecting to the grid, specifically around NPV rather
than total price.1.3. System Architecture
Generator: Two machines joined together, an ICE to provide torque and an electric generator to turn the torque into electricity which is then feed into the battery
PV: Converts the energy in sunlight into electricity
Battery: Stores excess electricity produced by the generator or PV panels and allows it to be used at a later time than when it was produced, normally night-time.2. REFERENCES
- Lilley, B; Satzow, A; Jones T; (2009) ‘A Value proposition for Distributed Energy In Australia’ CSIRO.
- WholeSaleSolar (2013). "Deep Cycle Battery Banks." Retrieved 10/10, 2013.
- Winaico (2013). "Technological leader: WINAICO QUANTUM." Retrieved 10/9/2013, 2013.
- Decker, K. D. (2009). "Small windmills put to the test." Retrieved 10/10, 2013, from http://www.lowtechmagazine.com/2009/04/small-windmills-test-results.html.
3.1. Some Advantages and Disadvantages of Islanded Microgrids
The reason we are looking at microgrids is because they offer advantages that are useful for a wide range of people in pursuit of various goals. Below is a sample of those advantages.
- Lower transmission losses
- Which leads to lower distribution costs
- Linked is the lower technical complexity
- Can be easily automated
- Is more disaster proof
- Can be set up to link to the grid and detach when necessary
- Lacks diseconomies of scale
Like anything, microgrids also have their disadvantages that make them unsuitable for all uses. Below is a list of some of those disadvantages.
- Can’t aggregate supply and demand from a large area
- Some energy sources aren’t suitable (large wind turbines)
- Requires battery or similar storage (instead of it being optional)
- The initial cost and F.I.T
- Lacks economies of scale
- Isn’t reliable unless well built
Since this system is designed to be owned and operated by the average Australian, its requirements are geared towards that sort of situation. Price is important, along with usability and it’s a functioning investment that needs to provide some sort of return.
The system shall be modifiable/future proof. Say any component can be replaced within another model with performance indicators within a 10% range
The total running costs of the system, including maintenance, fuel and repairs, shall be at the same price or lower than electricity from the grid
The set-up costs of the system shall be affordable ($10-30,000) and at or less than connecting to the grid
The technical skills require for running the system shall be readily obtainable by the clients
The system shall have a high reliability (about 99.9999%)
The system shall be islanded
The system shall provide hot water and electricity at an equal or greater rate than the client’s need, when they need it.
The system shall have a minimum of overnight storage (16Kwh) for winter months3.3. Introduction to Optimisation
Each of the various element falls under a different category: those that can be changed (variables), limits to the model (models), the outputs of the model and options.
- Size of the batteries
- Load management of the batteries
- Size of the PV panels
- Generator chosen and fuel choice
- Generator run time
- Cost of electricity
- The price has to be less than simply connecting to the grid
- Number of houses that can be connected
- Roof space
- Load (24KWhr)
- Spread of sunlight throughout the year
- System life cycle (20 years)
What’s being optimised
- Price of electricity needs to be as low as possible
- CO2 emissions need to be minimised
What can be changed to improve the design
- The batteries cycling is a separate optimisation between total capacity and lifespan
- Size of the PV’s and battery capacit
- CO2 emissions are directly proportional to generator run time
- Alternate generators and fuels will also affect CO2 emissions
- Size of the solar panels is inversely proportional to generator run time
- If there are enough solar panels, increased battery size can lower generator runtime
- Solar panels increase initial cost of the system
- Diesel use affects running cost
- The cost of fuel can fluctuate, while solar panels costs are locked in
- An increase in battery size increases cost
- Generator run time affects how many houses can be connected
- The cost of electricity is setup cost/lifetime plus maintenance costs
- Increase in storage (batteries or fuel tanks) reduces risk
- Risk here refers to the chances that no electricity will be provided when needed
- Battery load management affects battery lifespan
- Life cycle cost is simply the capital costs + maintenance cost divided by lifespan
- Load is serviced from the battery, there is no direct way for generated electricity to service load
PV power produced = #PV x PV% x Solar exposure x PV area
Excess Power = PV power – Load
Storage = Previous days storage + excess power: range 0<Storage<Max Storage
Fuel consumption = g/Kwh of diesel / generator %
Fuel use = Fuel consumption x deficit of storage
Fuel cost = total fuel use x fuel price
Capital cost = Battery price + generator + PV panel prices
Total cost = Capital cost + lifecycle x Fuel cost + replacement battery
Cost per Kwh = Total cost / (lifecycle x load x 365)
CO2 production = Total fuel use x CO2 production
Improvement over grid = Cost per Kwh – Comparison cost for grid
NPV = inflation^year*(cost/(1 + discount rate)^year)
- Two forms, one for grid electricity and the other for off grid
The generator is 90% efficient
-which means the engine has to produce 1.11 of the required load
We have unlimited roof space
Maintenance costs are negligible
No transmission loses
Batteries are 100% efficient, no electricity is lost by storage
The density of diesel is 0.832kg/L
Engine consumption is .204kg/Kwh
-therefore consumption is 0.245L/Kwh
The price of diesel is $1.5/L
CO2 production is 2.68/Kg
We are using the Winaico Quantum panels
-Efficiency is 17.46%
-Area is 1.663m2
-$290 per Panel
Meteorology data is for the botanic gardens, 2012 data
Projects lifecycle is 20 years
Load is 24kWh, spread evenly over a day
Assume batteries are fully charged at the beginning
Assume 27c kWh for comparison with the grid (based on recent electricity bill)
Any battery deficit is recharged by the generator that day
Components don’t lose performance over time
-batteries just have to be replaced
For a battery bank, I have chosen a specific model 12 Surrette 6v, 400 Ah S530 from whole sale solar
-Price $4497, load 28.8 Kwh and a lifespan of 10 years, warranty 7 years
-Assume that it will last all 10 years then die without losing performance
3.7 Flow chart of calculation
3.8 key data from optimised state
Optimal #PV: 43 panels
Set up cost: $10967.12
NPV improvement over grid: $38,892.70
NPV cost: $28,229
3.9 Interpretation of spread sheet & Discussion
If optimising for NPV, 43 solar panels is the optimal choice as it has the lowest NPV. Importantly, lowering the load decreases the advantage of going off grid, at 15Kwh it is better to stay connected to the grid. This suggests that adding more houses to the microgrid will further increase the advantage of going off grid, especially if those house have a low electricity usage. In that case this model simulates 3 houses of 8 Kwh load each fairly well and as long as the battery is about one days storage this approach is accurate.
Importantly, this model shows that reducing the battery size as much as possible is the optimal solution. But since calculations are done at a daily basis, when the battery gets below a day’s storage the model will be inaccurate because night-time electricity use isn’t accounted for. If calculations were done an hourly basis, then the day/night cycle can be taken into account and the variation of load over the day can be simulated, then battery storage can be lowered below a day’s storage without the model losing accuracy. This is the prime reason I choose an off the shelf battery, optimising the exact size by buying individual batteries isn’t an option supported by the current model.
The main component missing from the model is heat storage and hot water for space heating. This is actually a separate task that then informs the main model. It is also where the optimum number of attached houses (and thus final load) is likely to come from, more houses means a bigger heat storage tank but also more heat loses in moving the heat around. Having heat storage also allows the excess power produced in the summer months to be used rather than wasted, it can simply power an electric heating element or heat pump.
The heat storage would be modelled like the battery, except that storage lose is added, a certain amount is needed and if there isn’t enough the diesel generator is run for longer. In most cases, the diesel generator won’t be run that much more than it already is but when it does it will provide extra power. This power can again be used by electric heating elements to store heat for tomorrow. Heat for other uses can be added, but this is more complicated since hot water use for showers varies widely with households.