Method and instrumentation for sustainable energy load flow management system performing as resilient adaptive microgrid system

Abstract

Disclosed is a method and instrumentation for predictive and adaptive controllers devised to ensure uninterrupted operation of standalone electrical supply systems powered by sustainable energy sources. The device hereby referred to as SelfMaster is an expert system that manages the energy conversion, storage, and consumption in an isolated electric grid based on data collected during past and current operation of the system and predicted future states of the primary energy sources, storage level, and demand. The sustainable primary energy sources managed by SelfMaster may include, but not limited to, wind force, solar radiation, and biofuels. The energy storage system is a combination of batteries, hydrogen, biofuel, and hot water tanks. Electric demand consists of critical, non-critical, and deferrable loads identified according to the activities supported by the supply system.

Claims

1. A system and method of managing microgrids using Observer, Resource Estimator, Simulator, Scheduler, and Controller routines each considered as a separate virtual device created in the computer software as separate functions, comprising the steps of: using a processor of a computing device to check collected processed data to determine the status of the micro-grid; using the processor to determine if an electric battery bank is full, and if so then to command channelling excess energy to be stored in non-electrical form; using the processor to determine if a secondary energy storage system is full, and if so then to command diverting power into dummy loads; if the battery bank is not full, then using the processor to determine the likely available energy given weather data collected and weather forecasted in an observer routine, the step using a resource estimator routine to determine an amount of likely available energy required for an adequate level of electric storage, and then to: i) run a load scheduler which limits use of deferrable loads by sending signals from a controlling computer to turn off deferrable loads according to priorities until a simulation indicates an adequate level of electric storage; ii) then, if the level of electric storage is not indicated to be adequate, command on the use of the secondary energy storage system; then, iii) provide alarms and alerts to the system management system showing the use of the energy storage system; then, iv) check the rate of the energy storage system and continue to use the auxiliary energy system until the rate changes in order to meet the adequate level of electric storage; v) compare an amount of energy reserve in the secondary storage system and the observer routine to determine projected power coming into the system, and using the resource estimator routine to determine the amount of time the auxiliary storage system will provide the needed amount of energy and provide alerts and alarms to the system management system; using the processor to determine if the auxiliary energy storage system continues to discharge, and if so then, determine priorities of critical loads and begin to reduce the critical loads, maintaining the power management and alert system as the most critical load; then continue to send alarms and alerts to the management system; using the processor to determine if the auxiliary management system continues to discharge at an unacceptable rate, and if so then begin the final safe and orderly shut down of the system while maintaining a minimum of power and system management of the system; wherein each of the above steps utilizes at least one particular machine, said at least one particular machine comprising a computer and related industrial controls necessary to adjust power sources, storage systems and power using applications.

2. The method of claim 1 further comprising a step of collecting data to form a knowledge base for identification of actual component characteristics.

3. The method of claim 1 further comprising a step of the use of an expert system that learns the behavior of the components evaluating archived data.

4. The method of claim 1 further comprising steps of estimating the future states of the stochastic wind and solar resources, including temperature and precipitation, upon weather forecast data.

5. The method of claim 4, further comprising a step of using multiple sensors positioned to receive wind speed, solar radiation, sky cover, precipitation, and shading factors.

6. The method of claim 1 further comprising steps of assessing the amount of energy needed for critical loads in an energy system and considering the energy storage system itself as the most critical load to be maintained in order to protect the other critical loads.

7. The method of claim 1 further comprising a step of coordinating with a sponsored services routine whereby loads are provided to applications and users from accounts of those willing to pay for those loads.

8. The method of claim 1 further comprising a step of provisioning the microgrid with electromagnetic pulse protection.

9. The method of claim 1, wherein the non-electric form comprises at least one selected from the group consisting of: hydrogen, methane, other gaseous or liquid fuels, biofuel production, thermal, kinetic, or potential energy.

10. The method of claim 1, wherein the non-electric form comprises hot water.

11. The method of claim 1, wherein the non-electric form comprises a flywheel.

12. The method of claim 1, wherein the non-electric form comprises compressed air or pumped water.

Description

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING(S)

(1) FIG. 1: Outline of a sustainable standalone micro grid supplied by renewable energy sources and controlled by SelfMaster

(2) FIG. 2: Interaction of SelfMaster with a micro grid, power supply units, weather service provider, user, and other SelfMaster units.

(3) FIG. 3: Flowchart of energy management process performed by SelfMaster

(4) FIG. 4: Flowchart of the Observation Routine process

(5) FIG. 5: Data logging system

(6) FIG. 6: Flowchart of the resource estimation process

(7) FIG. 7: Flowchart of the simulation, scheduling, and control processes

DETAILED DESCRIPTION OF THE INVENTION

(8) The purpose of SelfMaster is to control load flow and storage in an isolated micro-grid isolated or potentially isolated from a larger grid as depicted in FIG. 1. The micro-grid is supplied by a group of wind turbines, solar PV arrays, and an optional auxiliary backup generator. The backup generator may be a combination of generators driven by a multi-fuel combustion engine, compressed air, flywheel, or fuel cell stack. Solar thermal collectors provide heating source. A hybrid storage system is a combination of primary and secondary storages. Primary storage is always a battery bank for short-term energy backup. Secondary storage may be hydrogen, biofuel, or compressed air tank, flywheel, or hot water tanks. The electric load is grouped as Critical Load, Non-critical Load, and Deferrable Load. Critical load consists of internal supply for sensors, computing devices, and controllers; communication, control, signalization (IT), and emergency lighting. Non-critical load is the equipment, appliances, and devices that are needed for a comfortable life or regular functions. Audiovisual devices, daytime artificial lighting, and part of the variable frequency drives (VFD) for HVAC circulation are examples to non-critical load. Deferrable loads can be supplied at a convenient time frame based on the energy state of the micro-grid. Due to the thermal capacity and longer time constant of the HVAC systems, part of the VFD load can be considered as deferrable load. Refrigeration, pumping for water storage, and device rechargers are also deferrable loads.

(9) SelfMaster is the central control unit that manages the load and storage based on current and estimated future states. Interaction of SelfMaster with an isolated micro-grid is shown in FIG. 2. In this figure, arrows show data flow and lines represent power connections. Data collection points are shown with a dot on the power lines. It is assumed that data collection sensors do not affect the voltage and current values on the power lines. [0013] The major components of SelfMaster are Observer, Resource Estimator, Simulator, Scheduler, and Controller routines. Each of these components is considered as a separate virtual device created in the computer software as separate functions. Data flow between these functions is shown in FIG. 2. An operator may interact with SelfMaster via a user interface to enter input data and monitor the system performance. User inputs consist of a list of activities to be scheduled, needed resources (such as space allocation, temperature, lighting, and equipment), and priority level of each activity. Data is logged both at the micro-grid and/or at a remote location. In addition, several SelfMaster units may communicate with each other to control a cluster of micro grids.

(10) Outline of the Operation

(11) The flowchart in FIG. 3 shows the overall operation of SelfMaster. Collected data is processed to determine the status of the micro-grid. If the battery bank is full, then the excess energy will be stored in non-electrical form such as (but not limited to) hydrogen, methane, other gaseous or liquid fuels, biofuel production, thermal (hot water), kinetic (flywheel), or potential energy (compressed air, pumped water).

(12) Weather forecast data for the following given number of days is automatically downloaded from a weather station (such as National Weather ServiceNWS) database every hour. The hourly generation, storage, and consumption values are estimated for a given time interval through real-time simulation based on the forecast information and user defined load profiles. The anticipated storage level is checked at every simulation and deferred load is scheduled to optimize the energy balance. If the storage level is expected to fall below a user defined critical level, then SelfMaster will start available auxiliary generation to charge the battery bank until the first upcoming simulation indicates an adequate level of electric storage.

(13) The forecast data relevant to the operation of SelfMaster are temperature (), surface wind speed (V), and percent sky cover (C). The computer program sends a SOAP request to the NDFD XML server through the Internet. The SOAP response received from the server is converted to a data table and stored in a file.

(14) Data acquisition hardware and software collect the DC voltage and current outputs and cell temperatures of the series connected PV modules. If the PV array is generating power, the DC output of the charge controllers and AC output values of the inverters are recorded simultaneously to compute the actual efficiencies and update the PV database.

(15) Similarly, a separate data acquisition system collects the output voltage, current, and frequency of the wind generators. If any of the wind turbines is generating power, then the DC output of the charge controllers and AC output values of the inverters are recorded simultaneously to compute the actual efficiencies and update the wind turbine (WT) database.

(16) Energy stored in the primary storage (battery bank) is determined by recording the actual charge and discharge amp-hours.

(17) The energy reserve available in the secondary storage is evaluated based on non-electrical quantities, such as temperature, pressure, volume of fuel, etc., depending on the type of energy stored. The amount of stored fuel is converted to electrical energy equivalent using the specific value of the stored substance such as hydrogen, methane, biomass, biodiesel, or anaerobic digestion products.

(18) Observation Routine

(19) The Observation Routine named hereafter observer receives inputs from sensors, a weather forecast service, and user interface. Sensors and data acquisition hardware collect electrical and non-electrical quantities such as voltages, currents, temperatures, liquid level, and pressures, etc. A local weather station on site provides current temperature, wind speed, sky cover, and precipitation data at the actual location. Forecast data is periodically downloaded from a weather station to record hourly temperature, wind speed, sky cover, and probability of snow precipitation for a given number of days. Collected data is stored in a local memory device and also sent to a remote storage device. In addition, the observer routine computes the actual efficiency of the generation units and updates the databases.

(20) Manufacturers usually give typical catalog specifications of wind turbines, solar PV modules, and converters based on factory tests and guaranteed rated values. However, the actual efficiency of these components depends on environmental conditions, aging, and possible faults or damages during service. The observer updates periodically the actual state of each power supply component for more accurate estimation. It also generates warning or alarm signals when critical generation issues occur. Data collected by sensors is logged in (both at the device and possibly in a remote location as part of a power management and network management system for the micro-grid) for reporting, troubleshooting, and future reference and, when appropriate, communicated to other elements of the micro-grid.

(21) FIG. 4 shows a flowchart explaining the observation process. The computer program downloads periodically weather forecast data for the following 105 hours from National Oceanic and Atmospheric Administration's (NOAA) National Weather Service NWS) database. NOAA updates the US National Digital Forecast Database (NDFD) every hour with forecasts data produced on a 3-hourly basis for up to three days ahead and on a 6-hourly basis up to six days ahead. NDFD provides gridded data for a location specified by either the postal (zip) code or GPS coordinates. Perez et al. (2010) present a validation of the NDFD short-term forecast. Kim and Augenbroe (2012) discuss the adequacy of NDFD forecasts for building automation and control processes. According to the evaluations presented in these publications, NDFD provides an acceptable level of accuracy up to a forecast horizon of six hours.

(22) While other forecast methods presented in Heinemann et al. (2006) and Perez et al. (2007) can provide more accurate forecasts beyond 6-hour horizon, SelfMaster uses the NDFD for the following reasons: Temperature, surface wind speed, sky cover index, and chance of snow precipitation are available in the database Service continuity and quality control is ensured by the US National Oceanic and Atmospheric Agency (NOAA) Service is updated every hour for a grid of 5-km spatial resolution for the contiguous US NDFD is an open source web service accessible to public through Simple Object Access Protocol (SOAP) in the Extensible Markup Language (XML) format.

(23) A schematic outline of the local data logging system is shown in FIG. 5. The data logger and sub-meters are off-the-shelf monitoring devices available on the market. The devices communicate with the data logger via RS485 connection using the MODBUS protocol. The serial data communication is described in the document MODBUS Organization, (2012). The data logger supplies data to a local computer and provides remote access through the Internet.

(24) Resource Estimation Routine

(25) The Resource Estimation Routine, named hereafter estimator, uses the database created by the observer. It computes the estimated power generation for the period of time covered by the weather forecast using four sources of data listed below. Forecasted wind speed, sky cover, and snow precipitation data downloaded from NDFD Actual generation data collected by the local data acquisition system Extraterrestrial global solar radiation data obtained from NREL SolPos software Characteristics of solar PV modules and wind turbines provided by the manufacturers

(26) The first step of the estimation process is to correlate the actual PV generation to the sky cover and snow precipitation history recorded over the last N.sub.p number of days. A reasonable default value of 30-day is selected for N.sub.p to record the seasonal variation of the solar path, average temperature, shading, snow, and dust cover, or any loss of energy due to equipment faults. Extraterrestrial global radiation given in W/m.sup.2 does not depend on the sky cover conditions. A number of methods to estimate the clear sky irradiation at a location on the earth surface were compared by Reno et al. (2012). Simple clear sky models only based on geometric calculations can be used in estimation of the global irradiation since the sky cover data already include atmospheric parameters considered in more complex models. Average errors of various models as a percentage of measured irradiance for 30 sites in the US are compared in FIG. 22 of Reno et al. (2012).

(27) Ineichen and Perez model proves to be reasonably adequate with 5% Root Mean Square Error (RMSE). Expression (1) below describes the Global Horizontal Irradiation (GHI) for the Ineichen and Perez model.
GHI=c.sub.g1.Math.I.sub.0.Math.cos(z).Math.e.sup.c.sup.g2.sup.M.sup.a.sup.[f.sup.h1.sup.+f.sup.h2.sup.(T.sup.L.sup.1)].Math.e.sup.0.01.Math.M.sup.a.sup.1.8(1)

(28) In this expression z is the zenith angle calculated for the location and time, I.sub.0 is the extraterrestrial normal incident irradiation, A.sub.m represents the air mass, and T.sub.L is the atmospheric (Linke) turbidity reformulated by P. Ineichen and R. Perez, (2002)
c.sub.g1=5.09.Math.e.sup.5h+0.868
c.sub.g2=3.92.Math.e.sup.5h+0.0387(2)

(29) In (2) h is the elevation.

(30) The Atmospheric Efficiency E.sub.a is defined here as the ratio of the actual DC power generated by the PV array and the DC power this array would produce for the GHI calculated for the given location and time. The atmospheric efficiency is zero at night. The E.sub.a value obtained at a given instant during daytime is a function of many factors such as cloudiness, clearness, water vapor content, and ozone layer thickness. The estimator first obtains a linear correlation of the computed atmospheric efficiency values and the recorded sky cover values at the observation times. As well as the atmospheric conditions, shading, dusting, and minor defects of the modules are included in this correlation.

(31) Resource Estimation Methodology

(32) The Resource Estimation Routine is outlined in FIG. 6. The purpose of this routine is to estimate available energy sources, electric storage, and energy capacity of stored fuel at the current time and for the following N.sub.f number of hours.

(33) The sky cover (cloudiness) index provided in the NDFD is not directly related to the solar irradiation received at the earth surface. In addition, microclimate, shading, dust cover, and aging affect the output power of PV modules.

(34) Simulation and Scheduling Routines

(35) FIG. 7 shows a flowchart for the simulation and scheduling, and control routines. The Simulation Routine, named hereafter simulator, receives actual power generation data and performance characteristics from the database generated by the observer. A Monte-Carlo simulation is performed to estimate the generated power P.sub.g and cumulative stored energy W.sub.s(t) over the forecast horizon of N.sub.f hours.

(36) An operator (user) enters planned activities by specifying the priority level, planned start and end times, light, heat, and equipment needed for each activity. The scheduling routine estimates the energy needed for the requested activities (W.sub.a) and tries to place them at the requested time slots on the schedule. The difference between available and needed energy at all instants is computed.
W.sub.e.sup.k=W.sub.g.sup.k+W.sub.s.sup.kW.sub.a.sup.k(3)

(37) If the simulations do not guarantee sufficient energy at all instants for the requested activities (i.e. W.sub.e.sup.k>0), the scheduling routine may shift deferrable loads to obtain an optimal load distribution. If the needed energy is still not available, the scheduler suggests better time frames available for the requested activities or recommends the user to reschedule or revise the request.

(38) The simulator and scheduler routines interact to find an optimal activity schedule that can be supplied by the available resources. If the iterations converge to the optimal load distribution over the forecast horizon, then the final schedule is forwarded to the controller, which sends signals to the switching hardware to turn on or off groups of critical, non-critical, and deferrable loads as well as activate the secondary storage or auxiliary generation units if needed.