Power generation control for remote cabins

Abstract

A system (1) including a mini-grid (101) for providing stand-alone power is provided. The mini-grid (101) comprises a plurality of appliances (10); a renewable source (20) for generating electricity f

Description

The present invention relates to the use of renewable energy to power a mini-grid. Preferably, but not exclusively, the present invention relates to the use of renewable energy to power a portable cabin. It is known in the art to use renewable energy sources to provide power to remote installations, for example, cabins. Renewable energy sources can provide intermittent electrical power. Therefore, such cabins will often include a battery for storing excess energy as it is generated. Such cabins will also often include a generator to generate electricity by the combustion of fossil fuels when the power provided by the renewable source and the charge stored in the battery is insufficient to meet demand. The present invention provides a system including a mini-grid for providing stand-alone power, the mini-grid comprising: a plurality of appliances; a renewable source for generating electricity from renewable energy; a generator for generating electricity from combustion; a battery; a controller connected to the renewable source, the generator, the battery, and the plurality of appliances, wherein: the controller is configured to control how power is supplied to the plurality of appliances from each of the renewable source, the generator, and the battery, using a control algorithm; the control algorithm involves: obtaining a renewable energy forecast; predicting an electrical demand resulting from the use of the plurality of appliances; and predicting an electrical supply from the renewable source based on the renewable energy forecast; and the control algorithm uses the prediction of electrical demand and the prediction of electrical supply to determine a desired state of charge of the battery for minimising the use of the generator. There is also provided a method of controlling a mini-grid for providing stand-alone power, the mini-grid comprising: a plurality of appliances; a renewable source for generating electricity from renewable energy; a generator for generating electricity from combustion; a battery; a controller connected to the renewable source, the generator, the battery, and the plurality of appliances, wherein the method comprises: obtaining a renewable energy forecast; predicting an electrical demand resulting from the use of the plurality of appliances; predicting an electrical supply from the renewable source based on the renewable energy forecast; determine a desired state of charge of the battery for minimising the use of the generator using the prediction of electrical demand and the prediction of electrical supply; and using the controller to achieve the desired state of charge of the battery by controlling the supply of power to the plurality of appliances from each of the renewable source, the generator, and the battery, and by controlling the supply of power to the battery from each of the renewable source, and the generator. The mini-grid preferably is, or includes, a cabin. The renewable source may be arranged to generate electricity from solar power or from wind energy. The renewable source and cabin may form a unitary construction. For example, the renewable source may be a photovoltaic device or wind turbine on the roof of the cabin. The plurality of appliances may include devices for heating and / or cooling and / or lighting the cabin, and in some embodiments also devices within the cabin such as computers. The generator may include an engine, such as an internal combustion engine, that combusts a fuel. The control algorithm may use the forecasts of the renewable energy to build mathematical models for the prediction of demand and electrical supply. The control algorithm may use these models to control the distribution of power from the renewable source and the generator to the appliances and to and from the battery. The control algorithm may do so in order to minimise the use of fuel. The mathematical model of the electrical supply may be based on a forecast obtained from a remote server. Calibration data may be captured using sensors forming part of the minigrid. The forecast obtained from the remote server may be updated to form a calibrated forecast using the calibration data. The mathematical model of the demand may be initialised based on data captured for similar mini-grids. The similar mini-grids may be identified by matching the mini-grids based on stored characteristics. The stored characteristics may include data including the intended and / or actual use of the mini-grid. The generator and renewable source may supply three-phase AC electricity. The appliances may be powered by single-phase AC electricity, with the three-phases powering different appliances. The demand prediction may involve predicting an unbalanced demand across the three phases, and the algorithm may take this into consideration when determining a desired state of charge of the battery. Such a system can provide a more efficient use of power, thereby reducing use of fuel. For a better understanding of the invention, and to show how the same may be put into effect, reference will now be made, by way of example only, to the accompanying drawings in which: Figure 1 shows a first embodiment of a system for providing stand-alone power in accordance with the invention; Figure 2 shows a second embodiment of a system for providing stand-alone power in accordance with the invention; Figure 3 shows a flowchart of a first embodiment of a control algorithm in accordance with the invention; Figure 4 shows a flowchart of a method for predicting the future electrical demand resulting from the use of a plurality of appliances; Figure 5 shows a flowchart of an alternative method for predicting the future electrical demand resulting from the use of a plurality of appliances; Figure 6 shows a method of creating a demand model; and Figure 7 shows a method of predicting the electrical supply available from the renewable source. A first embodiment of a system 1 for providing stand-alone power in accordance with the invention is shown in figure 1. The system 1 is arranged to provide stand-alone power because it is not connected to an electricity grid. The system 1 may be a mini-grid that is isolated from a mains power network. The system 1 comprises: a plurality of appliances 10; a renewable source 20; a generator 30; a battery 40; and a controller 50 (for example, a computer, processor, or similar device). The plurality of appliances 10 may include one or more of: lighting; air conditioning devices; heating devices; computers; pumps; and / or tools (for example, drills or cutting tools). The generator 30 may generate electricity by combusting fuel, such as a fossil fuel or hydrogen. The generator 30 may be / comprise an internal combustion engine. Multiple generators 30 may be provided to provide power for distribution by controller 50. In some cases, the generator 30 may comprise a plurality of generators. For example, the generator 30 may comprise a generator that combusts a fossil fuel and a generator that combusts hydrogen. Battery 40 may be a single battery or multiple batteries collectively forming a battery pack. In the preferred embodiment shown in figure 1, the system 1 comprises a mini-grid 101. In this example, the mini-grid 101 includes a cabin 100, preferably a portable cabin. The minigrid 101 does not need to be a cabin and could be another construction, such as a chemical plant or office building, etc. Preferably, the cabin 100 houses one or more of the plurality of appliances 10 (one or more of the plurality of appliances 10 may be devices external to the cabin 100), the battery 40, and the controller 50. Preferably, the renewable source 20 and the generator 30 are located outside of the cabin 100. The cabin 100 may, for example, be a portable cabin for accommodating workers at a construction site. The plurality of appliances 10 may include devices to enable the workers to carry out the construction task. The plurality of appliances 10 may include devices to control the environment of the cabin 100, such as air conditioning devices. Purely as an illustrative example, such a cabin 100 may be occupied during the day, and empty overnight. Over a longer time scale, the cabin 100 may be occupied Monday to Friday, and empty over the weekend. As such, the demand for electricity for the plurality of appliances 10 may be predictable. In a more complicated example, there may be a work schedule associated with the cabin 100. The work schedule may indicate the likely use of appliances 10 at certain times. For example, the work schedule may indicate that labour requiring the use of electrical devices would be carried out over a particular time period. There is a variety of potential demands for electricity that may result from the use of the plurality of appliances 10 of the cabin 100. Similarly, the supply of electricity from the renewable source 20 may be intermittent. The battery 40 can be used to provide electricity to meet the demand of the plurality of appliances 10 when the supply of electricity from the renewable source 20 is insufficient. In the event that the battery 40, together with the renewable source 20, cannot sufficiently meet the demand provided by the plurality of appliances 10, the generator 30 may be turned on to provide a further source of electricity. It is desirable to minimise the total use of the generator 30 in order to reduce fuel usage. It is also desirable to minimise the use of the generator 30 during periods of occupation of the cabin 100. For example, in various embodiments this can maintain air quality and / or reduce costs (e.g., for hydrogen powered generators). It is also less efficient to use the generator 30 intermittently (i.e. switching it on and off frequently) rather than continuously (i.e. using it once for a continuous duration). One approach in which the generator 30 would be used more than is necessary the if the generator 30 was used to fully charge the battery 40 overnight, following a first day, in preparation for a second day. In this scenario, it may be that the renewable source 20 provided more electricity on the second day than was demanded by the plurality of appliances 10. The extra energy provided by the renewable source 20 could have been used to charge the battery 40. However the battery 40, being fully charged, could not store the extra energy provided by the renewable source 20 in this case. On the other hand, there is a scenario in which the generator 30 was not used to fully charge the battery 40, overnight, following the first day, in preparation for the second day. In that scenario, it may be that the renewable source 20 provided less electricity on the second day than was demanded by the plurality of appliances 10. The battery 40 would be relied upon to provide the shortfall. If the renewable source 20 and the battery 40 together were unable to meet the demand of the plurality of appliances 10, then the generator 30 would need to be used during the use of the plurality of appliances 10, and so during the occupancy of the cabin 100. As will be appreciated from the example scenarios above, a prediction of demand of the plurality of appliances 10, and a prediction of supply of electricity from the renewable source 20 for the second day, can enable a desired state of charge of the battery 40 to be calculated at the end of the first day. The generator 30 may be used to charge the battery 40 to the desired state of charge overnight, using a single use of the generator 30, while the cabin 100 is unoccupied, leaving the battery 40 at an appropriate state of charge to start the second day such that the generator 30 is used as little as possible on the second day. The inventors have realised that that the ability to predict future demand for electricity for the plurality of appliances 10 and the ability to predict future availability of renewable energy for electricity generation using the renewable source 20, can enable a better use of battery 40, improving overall efficiency. In order to prolong the lifespan of the battery, it is desirable to avoid the battery 40 from reaching a state of charge lower than a safe threshold. The system 1 may optionally comprise a remote server 200. The remote server 200 may be in communication with one or more other mini-grids 110, 112, 114, 116. The other minigrids 110, 112, 114, 116 may be identical to, or similar to, the mini-grid 101 of system 1. The other mini-grids 110, 112, 114,116 may each comprise a plurality of appliances 10, a renewable source 20, a generator 30, a battery 40, a controller 50, and a communication device 60. The other mini-grids 110, 112, 114, 106 may also be cabins. The controller 50 is connected to the renewable source 20, the generator 30, the battery 40, and the plurality of appliances 10. The controller 50 may receive electrical power from one or more of the renewable source 20, the generator 30, and / or the battery 40. The controller 50 may provide electrical power to the plurality of appliances 10. The renewable source 20 includes at least one intermittent source of renewable energy, the supply of which may be forecast. The renewable source 20 may comprise a photovoltaic device. A photovoltaic device of the renewable source 20 may be mounted on the cabin 100. The photovoltaic device can generate electricity from sunlight. A solar irradiance forecast may be obtained, for example via the Internet, to predict and electrical supply from the photovoltaic device of the renewable source 20. In embodiments in which the renewable sources comprises one or more photovoltaic devices to generate electricity, the mini-grid preferably also comprises one or more solar irradiance sensors for generating irradiance data. The solar irradiance sensors are preferably located adjacent the photovoltaic devices such that the generated irradiance data is representative of the light impinging on the photovoltaic device. Alternatively, or in addition, the renewable source 20 may comprise a wind turbine. The wind turbine of the renewable source 20 may be mounted on the cabin 100. The wind turbine can generate electricity from wind. A wind forecast may be obtained, for example via the Internet, to predict and electrical supply from the wind turbine of the renewable source 20. In embodiments in which the renewable sources comprises one or more wind turbines to generate electricity, the mini-grid preferably also comprises one or more wind speed sensors for generating wind speed data. The wind speed sensors are preferably located adjacent the wind turbines such that the generated wind speed data is representative of the wind driving the wind turbine. Multiple photovoltaic devices and / or multiple wind turbines may be provided to form renewable source 20. Multiple photovoltaic devices and / or multiple wind turbines may provide electrical power for distribution by controller 50. The controller 50 may receive electrical power from one or more of the renewable source 20, the generator 30, and / or the battery 40. The controller 50 may supply electrical power to one or more of the plurality of appliances 10 and / or the battery 40. A second embodiment of a system 2 for providing stand-alone power in accordance with the invention is shown in figure 2. The second embodiment differs from the first embodiment in that the mini-grid 101 encompasses more than a single cabin 100. Further cabins of 100a, 100b, 100c are provided. Cabins 100a, 100b, 100c are physically connected to cabin 100. The controller 50 of cabin 100 controls the provision of electrical power to appliances 10a, 10b, 10c, respectively installed in cabins 100a, 100b, 100c. Preferably, a single renewable source 20 is provided to provide power for all of the cabins 100, 100a, 100b, 100c. The systems 1, 2 shown in figures 1 and 2 may operate with three-phase power. In which case, the controller 50, may incorporate an inverter. That is, one or more of the appliances 10 may be three-phase appliances 10, powered using three-phase power provided by the controller 50. Optionally, one or more of the appliances 10 may be single-phase AC appliances 10, powered using one phase of the power provided by the controller 50. Owing to the variety of possible appliances 10, and the possibility of different phases of the three-phase supply being used to power different appliances 10, 10a, 10b, 10c, it is likely that the three phases of power provided by the controller 50 will be unbalanced. As a result, extra power will be needed in times of unbalanced load, in the form of reactive compensation. Reactive compensation therefore increases the amount of power required to be provided by the battery 40 than would be expected from a simple prediction of demand from the appliances 10, 10a, 10b, 10c. Without taking this into account, it would be necessary to utilise the generator 30 to provide the power needed for reactive compensation if the state of charge of the battery 40 matched the expected demand of the appliances 10, 10a, 10b, 10c. As discussed below, the control algorithm may be configured to separately predicted an electrical demand on each phase of the three-phase power to be provided by the controller 50. The control algorithm can therefore predict an electrical demand resulting from the use of the plurality of appliances 10 more accurately, by taking into consideration the phases with which they will be powered. The control algorithm may be stored on the controller 50, or in a remote server 200. In some cases, the control algorithm may be distributed between the controller 50 and the remote server 200. The purpose of the control algorithm is to determine an appropriate state of charge of the battery 40 for minimising the use of the generator 30. The control algorithm is arranged to predict an electrical demand resulting from the use of the plurality of appliances 10. The control algorithm is arranged to predict an electrical supply from the renewable source 20. The prediction of the electrical supply may be based on a forecast of the available renewable energy, such as a weather forecast, cloud cover forecast, temperature forecast, solar radiance forecast and / or wind forecast. The forecast may be obtained from a third party, for example, via the Internet. The control algorithm monitors the state of charge on the battery 40. The control algorithm may determine a desired state of charge of the battery 40. The desired state of charge of the battery 40 they be determined based on a prediction of electrical demand from the use of the plurality of appliances 10 and based on a prediction of the electrical supply from the renewable source 20. Based on the control algorithm, the controller 50 may use the generator 30 to charge the battery 40. The control algorithm is arranged to use the prediction of electrical demand and the prediction of electrical supply to determine a desired state of charge of the battery 40 for minimising the use of the generator 30. For example, purely as an illustration, the control algorithm may be implemented on a daily basis. There may be defined a use period, during which the mini-grid 101 is expected to be used. For example, the use period may be the period in which the cabin 100 is expected to be occupied by workers, based on a schedule (stored in the controller 50 or upon the remote server 200). Following the use period, the control algorithm may calculate the desired state of charge for the battery 40 for starting the following use period. The controller 50 may use the generator 30 to increase the state of charge of the battery 40 to the desired level (if the current state of charge is below the desired state of charge). As shown in figure 3, a control algorithm is provided to control the distribution of power by the controller 50. The first step 400 of the control algorithm is to determine the current state of charge of the battery 40. In some embodiments, battery 40 may comprise a battery management system (not shown), and the state of charge of the battery 40 may be read from the battery management system. In other embodiments, the battery state of charge 40 may be calculated. The second step 410 the control algorithm is to predict the future electrical demand resulting from the use of the plurality of appliances 10. The third step 420 of the control algorithm is to predict an electrical supply available from the renewable source 20. The fourth step 430 is to use the prediction of electrical demand and the prediction of electrical supply to determine the desired state of charge of the battery 40. The order of the first, second, and third steps 400, 410, 420 is unimportant. The steps may be carried out in any order. Step 410 of the control algorithm may be carried out by the method shown in figure 4. As can be seen in figure 4, the first step 500 of predicting the future electrical demand resulting from the use of the plurality of appliances 10 in accordance with step 410 is to create a demand model. The demand model may be a model of expected demand by the plurality of appliances 10 for each of a plurality of sub-periods over a period of time. In some embodiments, the demand model may be a model of expected demand by each of the plurality of appliances 10 for each of a plurality of sub-periods over a period of time. For example, the demand model may be an annual model indicating, for each day of a calendar year the expected demand of electricity by the plurality of appliances 10. The demand model may be initialised in a crude manner. For example, the demand model may be initialised by providing an equal expected demand for each day of planned use of the mini-grid 101. In a more refined example, if the plurality of appliances 10 include heaters or air conditioning systems (for example, full heating or cooling a cabin 100), the demand model may be initialised based on temperature forecasts for the location of the mini-grid 101 (that is, anticipating an increased demand for electricity for heating on cold days and for cooling on warm days). Such a demand model could be refined, if multiple cabins 10, 10a, 10b, 10c are provided by taking into consideration different expected occupancy of the different cabins. A further refining of the demand model could take into consideration a schedule of tasks to be carried out by the users of the mini-grid 101, for example, by the occupants of a cabin 100. For example, if the plurality of appliances 10 includes machinery that is expected to be used in accordance with the timetable, this can be used to create the demand model. In step 510, the method involves predicting the future electrical demand using the demand model. This could be, for example, looking up the expected electrical demand for a given period of time. In step 520, the actual electrical demand for that period of time is measured. The measured electrical demand can be used to update the model in step 530. The demand model may be generated via regression from historic data. In some embodiments, the demand model comprises sub-models. The sub-models may include a trend model, a seasonal variation model, a cyclical model, and a irregular fluctuation model. The models may be determined independently from one another. Preferably, the sub-models are combined by addition to form the overall demand model. The demand model can be initialised based on historic data, and / or be initialised based on a user input (for example, based on an expected schedule of use of the mini-grid). The demand model can be updated based on new data (for example, measured electrical demand) by combining the new data with the historic data. Preferably, the new data is combined with the historic data based on the age of the historic data. For example, the new data may be weighted relative to the historic data when updating the demand model (in this way new data may have a greater influence over the demand model than older data). In some instances, the new data is combined with the historic data using an exponential window function. As an illustrative example, a demand model for a new mini-grid may be initialised based on a time series of historic demand data associated with a different mini-grid. After the new mini-grid is used for a period of time, the system will have captured a first set of captured demand data. A first new demand model may be calculated based on the weighted combination of the historic demand data and the first set of captured demand data, with the first set of captured demand data having larger weights. Subsequently, a second set of captured demand data may be captured by the system. A further new demand model may be calculated based on the weighted combination of the historic demand data, the first captured demand data, and the second captured demand data, with the first captured demand data having larger weights than the historic demand data, and the second captured demand data having larger weights than the first captured demand data. As explained above, in some embodiments, the systems 1,2 shown in figures 1 and 2 may operate with three-phase power. Figure 5 shows a method of predicting the future electrical demand resulting from the use of the plurality of appliances 10 in accordance with step 410. In which case, as shown in figure 5, the first step 600 of predicting the future electrical demand resulting from the use of the plurality of appliances 10 is to create a demand model for each phase of the three-phase supply. In this case, each demand model can be created in the same manner as described in connection with figure 4. The second step 610 of predicting the future electrical demand may include predicting the future electrical demand for each phase of the supply for a given period of time. The third step 620 may include predicting the total demand of electricity for the plurality of appliances 10 for the given period of time. Predicting the total demand may involve predicting the power factor correction and / or estimating the reactive power required to compensate for the unbalanced nature of the predicted demand on each of the three phases. The fourth step 630 of predicting the future electrical demand is measuring the actual electrical demand for each of the three phases for the given period of time. In step 640, the measured electrical demand for each of the three phases is used to update the demand model (for example, in the same way as set out above in connection with step 530). As noted in connection with figure 1, the mini-grid 101 may be connected to a remote server 200, and the remote server 200 may additionally be in communication with one or more other mini-grids 110, 112, 114, 116. It is possible to use historic use data from other mini-grids to assist with the creation of a demand model, for example in steps 500 and 600. Historic demand data may be stored, for example, in the remote server 200. In order that relevant historic demand data may be identified for use in the generation of a demand model for a mini-grid 100, it is advantageous that each of the mini-grids 110, 112, 114, 116 is characterised with relevant information that can be used to inform the applicability of historic demand data captured for that mini-grid to other mini-grids. Figure 6 shows a method of creating a demand model that may be used in accordance with steps 500 and 600. In step 700, one or more mini-grids 110, 112, 114, 116 are characterised. The mini-grids 110, 112, 114, 116 may be characterised based on an associated number of users (for example, the occupancy 100 of the cabins of each mini-grid). Additionally, or alternatively, the mini-grids 110, 112, 114, 116, if formed as one or more cabins, may be characterised based on the number of cabins and / or the number of rooms of the cabins. Additionally, or alternatively, the mini-grids 110, 112, 114, 116 may be characterised based on the selection of equipment forming the plurality of appliances 20. For example, minigrids 110 and 116 may include among the plurality of appliances 20 an air conditioning unit, whereas mini-grid 112 may include a heater, and mini-grid 114 may include a drill. Additionally, or alternatively, the mini-grids 110, 112, 114, 116 may be characterised based on their location. For example, mini-grids 110 and 112 may be located in a hot country, whereas mini-grids 114 and 116 may be located in a mild country. Additionally, or alternatively, the mini-grids 110, 112, 114, 116 may be characterised based on their use. For example, mini-grids 110 and 112 may be used predominantly for office work, whereas mini-grids 114 and 116 may be used predominantly for construction work. In step 710, the remote server 200 collects electrical demand data from each of the minigrids 110, 112, 114, 116. For example, this data may be collected during the normal use of each mini-grid. In step 720, the remote server 200 stores the historic demand data. The data may be stored in association with the identity of the mini-grid from which it was obtained and / or the characteristics of that mini-grid. In step 730, when initialising the demand model for a new mini-grid 100, the new mini-grid 100 may be characterised in the same way as the existing mini-grids 110, 112, 114, 116. In step 740, the demand model for the new grid may be initialised based on the stored historic demand data with the associated characteristics of the existing mini-grids 110, 112, 114, 116, and the characteristics of the new mini-grid 100. That is, the characteristics of the new mini-grid 100 can be used to identify similar existing mini-grids 110, 112, 114, 116. A single similar existing mini-grid may be identified, or a plurality of existing mini-grids may be identified, using a matching algorithm. The matching algorithm may involve one or more of: coarsened exact matching; nearest neighbour matching; propensity score matching; and / or sub-classification weighting. Figure 7 shows a method of predicting the electrical supply available from the renewable source 20 in accordance with step 410. As can be seen in figure 7, the first step 800 of predicting the electrical supply available from the renewable source 20 in accordance with step 410 is to determine the location of the mini-grid 101. In some cases, this may involve the determination of the location of the renewable source 20. However, the location of the mini-grid 101 will provide adequate accuracy for the forecasting of energy generation using the renewable source 20. The second step 810 involves obtaining forecast data from a forecast for the determined location. For example, if the renewable source 20 comprises a photovoltaic device, then a solar irradiance forecast may be obtained. If the renewable source 20 comprises a wind turbine, then a wind forecast may be obtained. Depending on the particular types of devices forming the renewable source 20 (this may include solar, wind, or other), one or more forecasts for each type of renewable energy may be obtained and combined to provide a forecast of expected electricity generated by the renewable source 20. In step 820, a predicted power model may be created of the expected power generation from the renewable source 20. For example, the predicted power model may be initialised based on the forecast. The predicted power model may be a model of expected electricity generated by the renewable source 20 for each of a plurality of sub-periods over a period of time. For example, the predicted power model may be an annual model indicating, for each day of a calendar year the expected generation of electricity by the renewable source 20. In step 830, the model is used to predict the power generated by the renewable source 20 for a given period of time. In step 840, the accuracy of the forecast for the given period of time may be measured using one or more sensors to provide sensor data. The one or more sensors may measure irradiance and / or wind speed. For example, if the renewable source 20 comprises a photovoltaic device, then the one or more sensors may include an irradiance sensor. If the renewable source 20 comprises a wind turbine, then the one or more sensors may include and anemometer. Optionally, in step 850, the actual power obtained by the renewable source 20 may be measured. A In step 860 the sensor data may be used to update the predicted power model. For example, the sensor data may be used to calibrate the forecast data. The calibrated forecast data may then be used to generate predicted forecast power data. The historic forecast data may be pooled with the predicted forecast power data, and regression may be used to create the predicted power model in step 860. That is, the control algorithm may involve regression of the combined historic forecast data and the predicted forecast power data. Preferably the predicted forecast of power data is combined with the historic forecast data based on the age of the historic forecast data. For example, predicted forecast power data may be weighted relative to the historic forecast data when updating the predictive power model (in this way new data may have a greater influence over the predicted power model than older data). In some instances, the predicted forecast power data is combined with the historic data using an exponential window function. When step 850 is present, this may also form part of the regression analysis. That is, the control algorithm may involve regression of the combination of the historic forecast data, the predicted forecast power data, and the measured obtained power data. In addition to, or instead of, using an exponential window function for weighting the predicted forecast power data relative to the historic forecast data, the measured obtained power data may also be weighted. For example, the measured obtained power data may have a higher weighting than the predicted forecast data or the historic forecast data. Moreover, historic measured obtained power data may be weighted based on its age. This may be done using an exponential window function.

Claims

1. A system including a mini-grid for providing stand-alone power, the mini-grid comprising: a plurality of appliances;a renewable source for generating electricity from renewable energy;a generator for generating electricity from combustion;a battery;a controller connected to the renewable source, the generator, the battery, and the plurality of appliances,wherein:the controller is configured to control how power is supplied to the plurality of appliances from each of the renewable source, the generator, and the battery, using a control algorithm;the control algorithm involves:obtaining a renewable energy forecast;predicting an electrical demand resulting from the use of the plurality of appliances; andpredicting an electrical supply from the renewable source based on the renewable energy forecast; andthe control algorithm uses the prediction of electrical demand and the prediction of electrical supply to determine a desired state of charge of the battery for minimising the use of the generator.

2. The system of claim 1, wherein the mini-grid comprises one or more portable cabins, the cabins being electrically connected and housing at least some of the plurality of appliances.

3. The system of claim 1 or claim 2, wherein:the control algorithm uses the prediction of electrical demand and the prediction of electrical supply to determine a desired state of charge of the battery for minimising the use of the generator by:determining the current state of charge of the battery; andpredicting the difference between the electrical demand and the electrical supply for a period of time following the present time; andthe controller has an operating mode in which it controls the generator to charge the battery to the desired state of charge.

4. The system of any preceding claim, wherein:the renewable source and generator are arranged to provide three-phase AC power;the controller supplies a first phase of the three-phase AC power to a first of the plurality of appliances;the controller supplies a second phase of the three-phase AC power to a second of the plurality of appliances;the controller supplies a third phase of the three-phase AC power to a third of the plurality of appliances;the control algorithm predicts the electrical demand by:separately predicting the electrical demand for the first, second, and third appliances; andusing the separately predicted demands to predict a power factor correction and / or an amount of reactive compensation required.

5. The system of any preceding claim, further comprising:a remote server; anda communication device connected to or integral with the controller for communicating with the remote server,wherein:the remote server stores historic usage data from other remote mini-grids; and the control algorithm predicts the electrical demand for the mini-grid based at least in part on the historic usage data.

6. The system of claim 5, wherein:the controller is configured to provide local usage data to the remote server via the communication device, the usage data representing the use of the plurality of appliances; andthe control algorithm predicts the electrical demand for the mini-grid based at least in part on the local usage data and the historic usage data.

7. The system of claim 6, wherein:the remote server stores data characterising the remote mini-grids; andthe control algorithm predicts the electrical demand for the mini-grid based at least in part on the local usage data and the occupancy and location of the remote minigrids, and the historic usage data and the occupancy and location of the remote mini-grids.

8. The system of claim 6, wherein:the mini-grid and the remote mini-grids each comprise one or more portable cabins; the remote server stores data representing the occupancy and location of the one or more portable cabins; andthe control algorithm predicts the electrical demand for the mini-grid based at least in part on the local usage data and the occupancy and location of the mini-grid, and the historic usage data and the occupancy and location of the remote mini-grids.

9. The system of any preceding claim, further comprising a renewable energy sensor for sensing the availability of renewable energy for the generation of electricity by the renewable source, wherein:the renewable energy sensor generates renewable energy sensor data; and the control algorithm calibrates the renewable energy forecast using the sensor renewable energy sensor data.

10. The system of any preceding claim, wherein:the controller measures the amount of electrical power obtained from the renewable source and thereby provides measured obtained power data;the control algorithm generates a forecast of the renewable energy based on a renewable energy model;the renewable energy model is initialised based on initial renewable energy data; andthe renewable energy model is updated based on the measured obtained power data.

11. The system of claim 10, wherein the renewable energy model is updated based on the measured obtained power data, by weighted the measured obtained power data depending on its age.

12. A method of controlling a mini-grid for providing stand-alone power, the mini-grid comprising:a plurality of appliances;a renewable source for generating electricity from renewable energy;a generator for generating electricity from combustion;a battery;a controller connected to the renewable source, the generator, the battery, and the plurality of appliances,wherein the method comprises:obtaining a renewable energy forecast;predicting an electrical demand resulting from the use of the plurality of appliances; predicting an electrical supply from the renewable source based on the renewable energy forecast;determine a desired state of charge of the battery for minimising the use of the generator using the prediction of electrical demand and the prediction of electrical supply; andusing the controller to achieve the desired state of charge of the battery by controlling the supply of power to the plurality of appliances from each of the renewable source, the generator, and the battery, and by controlling the supply of power to the battery from each of the renewable source, and the generator.

13. The method of claim 12, wherein the mini-grid comprises one or more portable cabins, the cabins being electrically connected and housing at least some of the plurality of appliances.

14. The method of claim 12 or claim 13, further comprising:using the prediction of electrical demand and the prediction of electrical supply to determine a desired state of charge of the battery for minimising the use of the generator by:determining the current state of charge of the battery; andpredicting the difference between the electrical demand and the electrical supply for a period of time after the present time; andcontrolling the generator to charge the battery to the desired state of charge.

15. The method of any one of claims 12 to 14, further comprising:provide three-phase AC power with one or both of the renewable source and generator;supplying a first phase of the three-phase AC power to a first of the plurality of appliances;supplying a second phase of the three-phase AC power to a second of the plurality of appliances;supplying a third phase of the three-phase AC power to a third of the plurality of appliances;predicting the electrical demand by:separately predicting the electrical demand for the first, second, and third appliances; andpredicting a power factor correction and / or an amount of reactive compensation required using the separately predicted demands.

16. The method of any one of claims 12 to 15, further comprising:a remote server; anda communication device connected to or integral with the controller for communicating with the remote server,wherein the method comprises:storing historic usage data from other remote mini-grids on the remote server; and predicting the electrical demand for the mini-grid based at least in part on the historic usage data.

17. The method of claim 16, further comprising:providing local usage data to the remote server via the communication device, the usage data representing the use of the plurality of appliances; andpredicting the electrical demand for the mini-grid based at least in part on the local usage data and the historic usage data.

18. The method of claim 17, further comprising:storing data characterising the remote mini-grids on the remote server; and predicting the electrical demand for the mini-grid based at least in part on the local usage data and the occupancy and location of the remote mini-grids, and the historic usage data and the occupancy and location of the remote mini-grids.

19. The method of claim 17, wherein the mini-grid and the remote mini-grids each comprise one or more portable cabins, and the method comprises:storing data representing the occupancy and location of the one or more portablecabins on the remote server; andpredicting the electrical demand for the mini-grid based at least in part on the local usage data and the occupancy and location of the mini-grid, and the historic usage data and the occupancy and location of the remote mini-grids.

20. The method of any one of claims 12 to 19, further comprising a renewable energy sensor for sensing the availability of renewable energy for the generation of electricity by the renewable source, the method further comprising:generating renewable energy sensor data with the renewable energy sensor; and calibrating the renewable energy forecast using the sensor renewable energy sensor data.

21. The method of any one of claims 12 to 20, wherein:measuring the amount of electrical power obtained from the renewable source to provide measured obtained power data;initialising the renewable energy model based on initial renewable energy data;updating the renewable energy model based on the measured obtained power data; andgenerating a forecast of the renewable energy based on a renewable energy model.

22. The method of claim 21, wherein the renewable energy model is updated based on the measured obtained power data, by weighted the measured obtained power data depending on its age.

Images