WO2021044132A1

WO2021044132A1 - Method and system for optimising battery usage

Info

Publication number: WO2021044132A1
Application number: PCT/GB2020/052089
Authority: WO
Inventors: Parisa AKABER; Stefan HAASS
Original assignee: Siemens Plc; The University Of Newcastle
Priority date: 2019-09-02
Filing date: 2020-09-01
Publication date: 2021-03-11
Also published as: GB2586654B; GB201912596D0; GB2586654A

Abstract

We describe a system and method for optimising battery usage, particularly within an energy storage system. The method may comprise measuring a set of variables for the at least one battery; selecting parameters for a degradation model which predicts degradation of the at least one battery; obtaining a degradation value for the battery using a predicted degradation value which is predicted using the degradation model and the selected parameters; obtaining historical data from one or more services to which the energy storage system is connectable; determining, using the degradation value and the historical data, an optimum state for the at least one battery for each of a plurality of time windows, and controlling the energy storage system based on the determined states for the at least one battery

Description

METHOD AND SYSTEM FOR OPTIMISING BATTERY USAGE

FIELD OF INVENTION

The present invention relates to a method and system for optimising battery usage, for example for lithium ion batteries within an energy storage system.

BACKGROUND OF INVENTION

Lithium ion batteries are increasingly being deployed in a variety of applications, including grid-scale power storage and in electric vehicles. For these various applications to perform optimally, a detailed understanding of the degradation of relevant life cycle battery metrics is essential. The relevant parameters may include the capacity and the resistance of the battery and their degradation is dependent on a number of factors. A generic physico chemical model is typically therefore not suitable to give reliable end of life (EOL) prediction or even more detailed state of health (SoH) information for specific batteries.

A paper entitled “Review of the remaining useful life prognostics of vehicle lithium-ion batteries using data-driven methodologies” by Wu et al published in Applied Sciences, vol 6, no 6, p166 May 2016 reviews various machine learning algorithms for predicting the remaining useful life (RUL) of vehicle lithium-ion batteries.

For grid-scale power usage systems, a contributing factor to the degradation is the use of the battery in an energy storage system which is supplying a particular market or service at a particular time. The different markets and services may have different technical requirements, different legal requirements and offer different revenue potential. For example, one service is termed the firm frequency response (FFR) and has various requirements, e.g. a minimum of 1MW response energy. Other example markets and services include the short-term operating reserve (STOR) service, the wholesale market (WSM) and the UK’s balancing mechanism (BM) service.

Therefore, there is a desire to provide an improved method and system for optimising battery usage, particularly for lithium ion batteries within an energy storage system.

SUMMARY OF INVENTION

According to a first aspect of the invention, there is provided a method for optimising usage of an energy storage system comprising at least one battery. The method comprises measuring a set of variables for the at least one battery; selecting parameters for a degradation model which predicts degradation of the at least one battery; obtaining a degradation value for the battery using a predicted degradation value which is predicted using the degradation model and the selected parameters; obtaining historical data from at least one service to which the energy storage system is connectable; determining, using the degradation value and the historical data, an optimum state for the at least one battery for each of a plurality of time windows, and controlling the energy storage system based on the determined states for the at least one battery.

The at least one service may comprise a plurality of services which may include both services and markets. For example, the plurality of services may comprise one or more of a balancing market (BM) service, a short-term operating reserve (STOR) service, a wholesale market (WSM) and a firm frequency response (FFR) service. The energy storage system may be connected to each service when a bid to supply the service has been accepted. The BM service and the WSM market may be considered to be short-term services because typically a bid to supply such services is made shortly, e.g. hours, before the energy storage system must be connected to the service to deliver the supply. The STOR service and the FFR service may be considered to be long-term services because bids are typically made and accepted a long-time, e.g. days, before the supply is required. Depending on the service, there may be one or both of a requirement to provide energy, e.g. to discharge energy from the at least one battery into the service or to remove energy, e.g. to discharge energy from the at least one battery into the service.

The optimum state may be one of discharge to one of the plurality of services, charge to one of the plurality of services or remain idle. In other words, the optimum state may be to connect to a service selected from the at least one service, either to charge or discharge thereto as appropriate. When there is only one service, e.g., the BM service which allows an energy storage system to both charge and/or discharge, the optimum states may be to connect to the BM service, either to charge or discharge thereto as appropriate.

Determining the optimum state may comprise determining a quantity of energy (e.g. volume) by which the at least one battery is to be charged in the charge state and/or a quantity of energy by which the at least one battery is to be discharged in the discharge state. The quantity of energy to be charged/discharged may be determined based on the degradation value.

Historical data may be obtained from a plurality of services. Determining the optimum state may comprise determining any boundary conditions which must be met at a start or end of each time window, e.g. when switching between services. Boundary conditions may comprise for example one or more of a required level of capacity at the start of a time window, a required level of capacity which must be maintained when charging a service.

The one or more services may comprise at least one short-term service and at least one long-term service. Determining the optimum state may comprise determining the optimum state for the at least one battery for the at least one long-term service for each of the plurality of time windows; determining whether there are any time windows in which the at least one battery is available and is not connected to the at least one long-term service; and when it is determined that there are available time windows, determining the optimum state for the at least one battery for the at least one short-term service in the available time windows. In other words, the method comprises a short-term optimiser step in which available gaps are utilised for the short-term markets or services. The method may thus be considered iterative.

Determining the optimum state may comprise determining an optimisation value for each of the one or more services for each of the plurality of time windows. The optimisation value may be determined independently for each of the at least one services. Where there is a plurality of services, the highest optimisation value for each service may be selected for each time window.

Determining the optimisation value may comprise determining a bidding strategy and a probability of success for the bidding strategy. A bidding strategy may include an offer of a volume of energy to be charged/discharged and may be associated with a bid price. For example, only bids having a success probability above a certain threshold, e.g. 60%, may be offered. The bidding strategy together with the degradation value may be used to determine the optimisation value. The optimisation value may be indicative of the profit margin. Developing bidding strategies may be an integral part of the BM, STOR and FFR services but not the WSM market. Bidding strategies in general use historical data from the specific market or service to predict a suitable bid price and a corresponding success probability. A battery operator may not be participating on the market with his own bids and offers, but may go through an energy supplier, e.g. when the battery is relatively small and/or short time in delivery. The contract with the energy supplier may be the WSM prices that are going to be paid to the battery operator or the WSM prices that the battery operator has to pay depending on whether the battery operator is discharging or charging. The WSM prices are as published by a third-party service, e.g. ELEXON and a forecast 3^rd party source may be used to estimate a day ahead. For example, for the STOR service, the FFR service and the BM service, the bidding strategy and success probability may be obtained from the historical data, e.g. analysis of events in the power grid and/or from the historical market information which is available.

The historical data may comprise prices and successful bids as a function of time (intraday, week-days/-ends, seasons). This historical data may be continuously updated with new data from the service to further optimize the optimisation value.

For example, for the BM service or the FFR service, the optimisation value may be determined using a Markov Decision process. For example, the set of actions l(s) for a BM service may be determined from

Where l(s) is the set of actions to be taken a is an action from the set of actions A(s) s is the current state at settlement index k=i s’ is the next state at settlement index k=i+1 from the set of states S

P_a(s, s’) is the success probability

R_a(s, s’) is the reward function

Y is a discount factor have a value between [0,1],

BatDeg is the battery degradation model including cycle counting algorithm with SoC(s’): state of charge as a function of state, T(s’): temperature as a function of state and rC: residual, normalised total capacity of the battery system at the initial state,

Capex: current capex costs for the battery system

EoL: end of life of the battery as a percentage value [0,100] of the battery as determined by manufacturer warranty

BoL: begin of life as a percentage value [0,100] of the battery system associated with capex value.

The success probability P_a(s, s’) may be determined from the input historical data, particularly the input historical market data. For example, the probability may be a function of the market price and quantity. The probability of success may also be a function of the energy storage characteristics, namely how much energy a particular system could supply for that settlement period. Thus, the success probability may be based on the degradation value. The revenue function R_a(s, s’) may be generated from the volume multiplied by the unit price. For the WSM market, the optimisation value may be calculated using an optimization engine which finds the minimum of an objective function F. The optimisation engine may comprise boundary conditions. The objective function F may be defined as: d d _ 100 x Capex

F = — —SoCd x CostD + — SoCc x CostC + BatDeg(SoC(t), T(t), rC ) x - - dt dt ( EoL — BoL) with

SoCd: state of charge as a function of time during discharging SoCc: state of charge as a function of time during charging

CostD: electricity price as determined from day ahead market forecast for selling electricity (e.g. input market data)

CostC: electricity cost as determined from day ahead market forecast for buying electricity (e.g. input market data)

BatDeg: Battery degradation model including cycle counting algorithm with SoC(t): state of charge as a function of time T(t): temperature as a function of time rC: residual, normalised total capacity of the battery system at the begin of the day Capex: current capex costs for the battery system (not including inverter, transformer, etc..) EoL: end of life of the battery as determined by manufacturer warranty (e.g. 70% of original total capacity) as % [0, 100]

BoL: begin of life total of battery system associated with capex value (usually 100%) as % [0, 100]

The degradation model may comprise a calendar ageing component and a cycling ageing component. Obtaining the degradation value may comprise obtaining an estimated degradation value for the battery using the set of measured parameters; and outputting a degradation value based on the estimated and predicted degradation values.

By including separate calendar and cycling ageing components, it is possible to model the effect of time and usage on degradation separately. The predicted, estimated and final degradation values may be a value for the current capacity of the battery. The estimated degradation value which is obtained from the variables may be obtained from live measurements of the battery.

The method may further comprise updating the measurements of the set of measured variables for the battery; predicting an updated predicted degradation value for the battery using the degradation model and the updated degradation model parameters; obtaining an updated estimated degradation value for the battery using the updated set of measured variables; repeating the updating of the parameters for the degradation model based on the updated estimated and updated predicted degradation values and outputting an updated final degradation value based on the updated predicted and estimated degradation values. The updating of the measurements may be done in real-time whereby the measurements are live measurements of the battery. In other words, the method may be iterative and may be repeated at multiple time intervals, both to update the parameters for the degradation model and to generate an up-to-date output value using the updated degradation values.

The measured set of variables may comprise at least one of current, voltage, state of charge, depth of discharge, temperature, number of cycles, CP-rate, minimum power out, maximum power out, maximum temperature, minimum temperature, maximum cell voltage balance, minimum and maximum SoC.

These may be measured using any suitable technique. When updating the measurements, a sub-set of the variables may be remeasured. By measuring a smaller number of variables, the updates may be done in real-time.

The parameters for the degradation model may be selected based on the characteristics of the battery and at least some of the original measured set of variables. The characteristics may comprise at least one of a manufacturer of the battery and chemistry of the battery.

The chemistry of the battery may represent the composition of the chemicals within the battery and example chemistries include lithium iron phosphate (LFP), lithium nickel manganese cobalt oxide (NMC), lithium titanate oxide (LTO), lithium cobalt oxide (LCO), lithium manganese oxide (LMO), lithium nickel cobalt aluminium oxide (NCA). These selected parameters may be termed the initial or starting parameters. The parameters may be selected from a stored set of parameters, e.g. a plurality of constants arranged in a look up table against the appropriate manufacturer and chemistry.

The calendar ageing component which may also be termed a calendar ageing equation may be a function of the variables: state of charge, temperature and time. For example, the calendar ageing component may be defined using equation 1 below:

where SoC is state of charge, T is temperature, t is time, R is the gas constant in KJmol ¹ (8.314....E-03), E_AI is the activation energy in KJmol ¹ K ¹, bi, ai and bi are the parameters (which are selected or updated) bi, ai and bi are dimensionless fitting parameters.

The cycling ageing component may be a function of different variables to that of the calendar ageing component. Some of the variables may overlap. The cycling ageing component which may also be termed a cycling ageing equation may be a function of the variables: state of charge, depth of discharge, constant power (discharge/charge) rate, equivalent full cycles and temperature. The cycling ageing component may be defined using equation 2 below:

where EFC is equivalent full cycles, SoC is state of charge, DoD is depth of discharge, CP rate is constant power (discharge/charge) rate, T is temperature, E_A2 is the activation energy in KJmol ¹ K ¹, b2, a2, b2, C2, and d2are the parameters (which are selected or updated). b2, a2, b2, C2, and d2 are dimensionless fitting parameters.

Updating the parameters for the degradation model based on the updated first and second degradation values may comprise using a Kalman Filter, e.g. an extended Kalman Filter. Outputting the final degradation value may comprise outputting a weighted sum of the estimated degradation value and the predicted degradation value. The final degradation value (and the updated final degradation value where appropriate) may be determined using a Kalman Filter. A dual Kalman Filter may be used to both update the parameters and output the final degradation value.

As set out above, the parameters may be selected from stored data such as a look-table.

The method may further comprise collecting data relating to the degradation of a plurality of batteries. The degradation model may be generated using the collected data. The parameters may be generated as fitting parameters.

According to another aspect of the invention, there is provided a (non-transitory) computer readable medium carrying processor control code which when implemented in a system (e.g. a battery analyser) causes the system to carry out the method described above.

Another aspect of the present invention is a system for predicting battery degradation. The system may comprise a processor which is configured to carry out the method described above. The system may also comprise one or more sensors for collecting data from a battery. For example, the one or more sensors may include a voltage meter for measuring the voltage of the battery. The one or more sensors may include an ammeter for measuring the current of the battery. The system may also comprise a user interface which is configured to display the output from the processor.

BRIEF DESCRIPTION OF THE DRAWINGS

The above mentioned attributes and other features and advantages of this invention and the manner of attaining them will become more apparent and the invention itself will be better understood by reference to the following description of embodiments of the invention taken in conjunction with the accompanying drawings, wherein

FIG. 1 shows a flowchart of a method implementing according to one aspect of the invention;

FIG. 2 is a graph of measured State of Charge (SoC) against time;

FIG. 3 is a graph of predicted normalised total capacity against time;

FIG. 4 is a graph of predicted normalised total capacity against time for a segment of FIG. 3;

FIG. 5 shows a flowchart of a method implemented in conjunction with the method of FIG. 1 ;

FIG. 6 is a graph of normalised capacity against time for a battery;

FIG. 7 is an example of a graph plotting the correlation in the dimensionless fit parameters ai and bi with the graph shown in FIG 6;

FIG. 8 is a flowchart for obtaining a set of actions for a battery system;

FIG. 9 is an example of predicted data for a plurality of markets and services which may be determined in the method of FIG. 8;

FIG. 10 is a state diagram for use in a model within the method of FIG. 9 illustrating the state of the battery at different indices k=1 , 2, ... , N and actions to transition from one state to the next state; FIG. 11 is a state diagram for use in a different model within the method of FIG. 9;

FIG. 12 plots profit against settlement period which may be determined in the method of FIG.

8;

FIG. 13 plots commitment to provide a supply against time which may be determined in the method of FIG. 9

FIG. 14 shows a schematic block diagram of a system which can be used to carry out the methods above.

DETAILED DESCRIPTION OF INVENTION

Figure 1 shows a flowchart for analysing battery performance. In a first step, the initial characteristics of the battery are obtained (S100). The initial characteristics may include the manufacturer of the battery and example manufacturers are LG Chem, Samsung, Toshiba, SK Innovation or Sony. The initial characteristics may also include the chemistry of the battery and example chemistries include lithium iron phosphate (LFP), lithium nickel manganese cobalt oxide (NMC), lithium titanate oxide (LTO), lithium cobalt oxide (LCO), lithium manganese oxide (LMO), and lithium nickel cobalt aluminium oxide (NCA).

The next step is to set the parameters which are to be used in the degradation model (S102). Initially, the start parameters may be optimised for the manufacturer and the chemistry of the battery based on measurements which were taken under laboratory conditions as described in more detail in relation to Figure 5.

Once the model parameters are set, various variables may then be measured (S104). It will be appreciated that the measurements may be taken simultaneously with obtaining the battery characteristics. These measurements may be termed live data because they are captured in real-time. These measured variables may include some or all of:

• Current (ampere),

• Voltage (volt),

• state of charge (SoC) which has a floating value between [0,1] and represents the remaining charge inside the battery relative to its current total capacity; a value of 1 is a “full” battery and 0 is an “empty” battery; • depth of discharge (DoD) which has a floating value between [0,1] and represents the absolute difference in the minimum and maximum state of charge of a given semi-charge or discharge cycle,

• temperature of the interior of the battery (which has a floating value in Kelvin) - may be estimated or measured,

• equivalent full cycles (EFC) which has a floating value and is a measure of the amount of charge from both charging/discharging divided by the associated total capacity of the battery;

• constant power (charge/discharge) rate (CP-rate) which has a floating value and represents the ratio between current total capacity (A_h) divided by (charge/discharge) current [A] and

• time t in days.

It is noted that SoC and DoD may also be expressed as a percentage between [0,100] but the models defined below use a floating value of [0,1]

For example, Figure 2 shows how SoC (%) may vary over time for a battery which is being monitored. The chart below shows examples of values of the variables which may also be captured at a point in time for the specific battery:

The variables may be measured or determined using standard techniques. It is noted that SoC for the cycling ageing component described below may be the average SoC because the SoC is not constant during a semi-cycle. These measurements are used to obtain a value for the capacity of the battery which is based on the measurements and may thus be termed a measured capacity (step S106).

The measured capacity may be an indication of the state of health (SoH) of the battery. The value may be obtained in any suitable way, e.g. using a C-estimation algorithm such as using the equation below:

where z ) is the battery cell SoC at time t2, z(h) is the battery cell SoC at time h, Q is the battery cell total capacity in ampere-hours, i(t) is the battery cell current at time t in amperes, h is a unitless efficiency factor which may take on different values depending on whether the current is positive or negative and time is measured in seconds. The factor of 3600 converts seconds to hours.

Suitable variations of the equation above are set out in “Recursive approximate weighted total least squares estimation of battery cell total capacity” by Plett published in Journal of Power Sources 196 (2011) 2319-23331.

As shown in Figure 1 , at the same time as the measured capacity is being obtained, a prediction for the capacity may also be obtained using a degradation model (step S106).

This predicted value for the capacity may be based on the model parameters which were set in step S102 and the variables which were measured in step S104 and may be termed a predicted capacity. Like the measured capacity, the predicted capacity may be an indication of the state of health (SoH) of the battery. It will be appreciated that it is optional to simultaneously obtain the measured and predicted capacity and the predicted capacity may alternatively be obtained after or before the measured capacity.

The degradation model may comprise two components: a first component which models degradation of the battery over time and which may be termed a calendar ageing component and a second component which models degradation of the battery resulting from the number of cycles through which the battery has cycled and which may be termed a cycling ageing component. Both components may model the physico-chemical basics of battery degradation but may contain different parameters such as state of charge (SoC), depth of discharge (DoD), temperature (T), time (t), constant power (discharge/charge) rate (CP rate) and equivalent full cycles (EFC). As explained in more detail below, each component may be an empirical model which comprises a set of fitting constants (i.e. parameters) for each of the variables which are included in the component. When using the degradation model, the measured SoC profile (such as that shown in Figure 2) may be segmented into parametrised semi-cycles and time periods of calendar ageing (with no discharging/charging). The separate segments Si, S2_, ... SN are indicated on Figure 2 and the end points of each segment represent changes in the trend for the SoC value.

For example, in the first semi-cycle, the SoC is gradually increasing but in the second semi cycle, the SoC is constant in value. The parametrised semi-cycles may be input into either the calendar or cycling ageing components to provide a predicted value for the change in total capacity (AC or dC) for each semi-cycle.

For example, the change in the calendar ageing component in a time interval At of the capacity may be defined using the equation below which incorporates equation 1 above:

with

where C_s is the total normalised total capacity before the calendar ageing event, SoC is state of charge, T is temperature, t is time, R is the gas constant in KJmol ¹ (8.314....E-03), E_AI is the activation energy in KJmol ¹ K ¹, bi, ai, and bi, are fitting parameters which are selected based on the battery cell chemistry and manufacturer identified in the initial method step.

The determination of the fitting parameters is described in more detail below.

For example, the change in the cycling ageing component of the capacity for an EFC of AEFC may be defined using the equation below which incorporates equation 2 above:

with

where C_s is the total normalised total capacity before the cycling ageing event, EFC is equivalent full cycles, SoC is state of charge, DoD is depth of discharge, CP rate is constant power (discharge/charge) rate, T is temperature, E_A2 is the activation energy in KJmol ¹K¹, b2, a2, b2, C2, and d2, are fitting parameters which are selected based on the battery cell chemistry and manufacturer identified in the initial method step. The determination of the parameters which include the fitting parameters is described in more detail below.

The output predicted capacity C_p may be predicted by iteratively subtracting from the original value for the capacity Co all the predicted changes in total capacity (dC,) for each of n semi cycles, e.g.

Cp = C₀ - dCi

In the example above, the two components both include the variables temperature and state of charge which can be readily measured. Each of the components also includes one or more additional variables which are specific to that component, e.g. time for the calendar ageing component and equivalent full cycles (EFC), depth of discharge (DoD) and constant power (discharge/charge) rate (CP rate) for the cycling ageing component.

The next step is then to compare the measured capacity with the predicted capacity (step S110). The comparison may be performed using a Kalman Filter, e.g. an extended Kalman Filter. A suitable Kalman Filter is described in detail in “Extended Kalman filtering for battery management systems of LiPB-based HEV battery packs” by Plett published in Journal of Power Sources 134 (2004) 262 to 292. As described with more detail in relation to Figure 3, the comparison step may lead to an output value for the current capacity which is based on both the estimated and the predicted values (step S112).

The comparison step may also be used to output updated values of the parameters in the degradation model (step S114). These updated values may be used as the new set of model parameters in step S102 so the process is iterative. The other steps are then repeated to generate a new output value of the current capacity of the battery and a new updated set of parameters which are based on the new measured battery variables.

The updated parameters may also be used together with an input planned usage profile (step S116) to obtain a future capacity of the battery using the updated degradation model and the planned usage profile (step S118). As described above, the method combines an empirical model (namely the degradation model) with an iterative learning algorithm to output a value (e.g. capacity) which is indicative of the state of health (SoH) of the battery and to update or adjust the model based on measured values.

A dual Kalman filter comprises a first Kalman filter and a second Kalman filter may be used for the comparison step of Figure 1. A first Kalman filter may apply a time update step which inputs the predicted states and the current and outputs the estimated states and estimated parameters. State estimation is done using the underlying degradation model and experimentally verified estimations for measurement uncertainties and errors in the shape of a Gaussian covariance matrix. The first Kalman filter may also apply a measurement update step in which the estimated capacity value and its associated estimated error and noise together with predicted capacity value and its associated error are input. The estimation algorithm described above may be used to estimate the error and noise of the estimated capacity value. Both the estimated and the predicted capacity value are combined using a weighted average which is dependent on the covariance matrices of each measurement/model error to give an output for the current capacity value (SoH).

A second Kalman filter may apply a time update step which inputs the predicted parameters and outputs the estimated parameters. These estimated parameters are used in the measurement update step of the first Kalman filter. These estimated parameters are also used in a measurement update step of the second Kalman filter together with the estimated parameters from the time update step of the first Kalman filter. Thus, there is an iterative adjustment of the model parameters.

Figure 3 plots the variation in the output predicted capacity over a long time period, e.g. between 0 to 275 days. The long-term prediction may be determined using the planned usage profile and the iteratively adjusted degradation model. Figure 4 plots the variation in the output predicted capacity over a shorter time period, e.g. on day 50, to give a more fine grained analysis.

Figure 5 is a flowchart illustrating how the empirical model may be determined. There is a first “acquisition phase” in which data from a plurality of batteries is obtained. In an initial step, the batteries are set-up (step S200) so that the degradation of the battery over time may be measured. A plurality of batteries from different manufacturers and different chemistry types may be set-up. The effect (if any) of the conditions such as temperature, current etc. is also acquired by exposing different batteries to different conditions. The data comprises total capacity over time and an example set of values is plotted in Figure 6. The time scale is hundreds and possibly thousands of days. Each data set comprises one specific set of ageing parameters which is determined based on the component of the model. For the calendar ageing component, the data set comprises values for capacity and time with one of SoC or T at different fixed values. Merely, as an example, if the data shown in Figure 6 were collected at a fixed temperature of 25 degrees Celsius, the data set would also need to include the value for SoC at each value for capacity and time. For the cycling ageing component, the data set comprises values for capacity and time with one of SoC, DoD, CP-rate and T at different fixed values.

A sufficiently large data set is required for the model estimation step. Merely as an example, there may need to be data collected for 3 different temperatures and 9 values of SoC for the calendar ageing component to provide a sufficiently large data set. Similarly, there may need to be data collected for 2 to 3 different temperatures, 2 to 3 different values of SoC, 2 to 3 different values of DoD, 2 to 3 different values of CP-rate which gives overall at least 16 sets of data for the cycling ageing component. Accordingly, the next step is to obtain values for a plurality of battery variables at a plurality of intervals (step S202). For example, the variables may include current, voltage, SoC, DoD, temperature, CP-rate. The variables may be measured using standard techniques. The measurements may be taken at regular intervals, e.g. each week or each day when considering calendar ageing or after a certain number of cycles when considering cycling ageing.

In addition to measurements, other values may be identified (step S204). For example, a capacity value which is indicative of the state of health (SoH) of each battery cell may be calculated at each interval. All the values which have been obtained, through measurement or calculation, are then stored (step S206). For example, the values may be stored in a measurement grid and/or may be plotted in graphs. Figure 6 is an example of a graph plotting the change in capacity value over time. The measurements points are indicated by crosses and a fit curve is generated by the model (+R² value = 0.96279).

Once the data has been gathered, it is used to generate the model. This includes fitting the model to the stored values (step S208). The calendar ageing equation may be defined initially as:

Similarly, the cycling ageing equation may be defined initially as:

Such equations have a low number of parameters and are relatively easy to handle but nevertheless give a good fit result. For both simplified equations, the fits of the measured data indicate that bi and b2 can be considered constant for all the measured data. Accordingly, the dependency of the capacity on the measured data is expressed by ai and 02 which are functions of the relevant variables. This reduces cross-correlations between the parameters ai, 02, bi and b2. In other words: bi = constant; b₂ = constant (the constant values may be same or different); ch=f{SOC, T) a ₂=f(SOC, T, CP rate, DoD )

An empirical approach based on studying the available data is then used to determine the functions which fully define ai and 02. Various techniques can be used to fit the data to the equations. For example, the well-known least square approach may be used. The resulting functions are shown below:

The definitions for the terms in the equations are the same as those above. The underlying structure of each equation is motivated on the one hand by the physico-chemcial considerations (e.g. in the use of the exponential Arrhenius term for temperature dependency) as well as observed dependencies of degradation with measured values in laboratory experiments. The equations are relatively simplistic but sufficiently accurate representations of the chosen dependencies. Concerning the number of cycles, the equations may only be valid in a certain range of capacity which should include the end of life (EOL) capacity value. The parameters can be considered independent of EFC or time.

Figure 7 is an example of a graph plotting the correlation in the dimensionless fitting constants ai and bi with the graph shown in Figure 6. The values of ai and bi having a correlation value of 1 at R²=0.96279 are selected. Other graphs may also be plotted for different batteries from different manufacturers. The data may also be presented in any suitable format, e.g. in a look-up table. As explained above, these initial values of the constants which are determined from historic data are updated for the specific battery being modelled using the measured variables. Merely as an example, the following table provides the range of for the fitting parameters for different chemistries and manufacturers.

Figure 8 is a flowchart illustrating how the degradation model described above may be used in a method to operate a battery within a battery energy storage system. For example, the method may be used to determine which market or service is to be supplied by the battery energy storage system and where it is appropriate what is being supplied to the market or service. As detailed below, in certain markets and services, the battery energy storage system may supply energy into the markets or services, e.g. by discharging the battery system but in other markets or services, the battery energy storage may supply energy or may remove energy, e.g. by charging the battery system.

In an initial phase, which may be determined a data collection phase, the current battery data may be obtained (step S900). This battery data may include measured data, e.g. SoC and temperature and other data such as the SoH. The SoH may be obtained using the method described above and thus all the variables which are listed above may be measured. In addition the battery data may include some or all of minimum and maximum power out, minimum and maximum battery cell temperature, maximum cell voltage balance, minimum and maximum SoC (when not [0,100]).

In addition to the current battery data, historical data for the battery may also be obtained (step S904). This historical data may include the data previously obtained for the battery at earlier time indexes together with details of the markets and services (if any) the battery has previously supplied. The historical data may also comprise the historical data for the markets and/or services, e.g. the supply which was provided by the battery storage system, including volumes and prices.

In addition to the historical data, the market data for the current and future time frame may be obtained (step S902). The market data may include the requirements for future energy level storage, the timing of these requirements and the revenue to be generated by providing battery systems to supply these requirements. Example markets and services include the firm frequency response (FFR) service, the short term operating reserve (STOR) service, the wholesale market (WSM) and the balancing mechanism (BC) service. Once all the necessary data has been gathered, an estimation of the profit for each market or service may be made (step S906). The market data, battery data and historical data are shown as being obtained at the same time but it will be appreciated that these steps may be done in any order.

Figure 9 shows an example of the profit which may be obtained from each market or service for each time index. Merely as an example, the time index is 48 settlement periods (time windows), that is 1 day but it will be appreciated that it may be a different time interval. As shown, the profit may vary from approximately £4 to £15 per megawatt for the FFR service with different time bands during a day having similar profits. The profit is determined based on the expected revenue from providing the service or trading in the market and any costs such as battery system degradation costs and other operational costs. Accordingly, the profit may also be negative in some time periods as shown in the WSM graph.

A different model may be appropriate for determining the profitability of each market or service. For example, the balancing mechanism (BM) service is a tool used by the National Grid to balance electricity supply and demand close to real time. The service works on settlement periods of half-an-hour. Participants known as balancing mechanism units (BMUs) which include generators, energy storage systems, aggregators submit bids and offers to increase/decrease their generation/consumption one hour prior to real time. These offers and bids are accepted in a way that settles any imbalance (difference between generation and consumption) in the most economical way possible with a focus on energy storage participation. The BM in the UK is just one example of such a service and other countries may have similar services to which the method below may be adapted.

Figure 10 is a simplified state diagram illustrating some of the possible states, s, s¹ , s², s³, s¹², s²², s²³ at time indexes k = 1, 2, ... N for a battery energy storage system which may be used in a BM. It will be appreciated that for compactness, only part of the state diagram is shown. At each index and from each state, three actions are possible as described in more detail below. Each state is represented by a set of battery data, e.g. measured data such as one or more of SoC and temperature and determined data such as the SoH. The SoH may be determined as described above and thus other battery data may also need to be measured. The data for the current state S is obtained at the starting time index k=1. Once the current state is known, there are three possible actions for controlling the battery: charge, discharge and idle. If the action “idle” is taken, the battery data will remain unchanged but the state will change to s² and the time index to k=2. If either of the actions “charge” or “discharge” are taken, the battery data will change.

In the example, shown in Figure 10, the action is for the battery to remain idle as the time index increases to k=2. At this time index, there are again three actions for progressing to the next time index of k=3. The action “idle” is included to allow the time index to be changed which is critical for the method of determining the actions to be taken with respect to the BM described below.

It will be appreciated that not all bids are accepted. Thus, each stage of this tree would be generated automatically from all possible states the system could end up in based on taking all possible actions, including charging-bid accepted, discharging-offer accepted, and staying idle because no bid offer was submitted.

Once all the data has been obtained (in whatever order), the next step is to obtain the probability of taking each of the actions which move the battery from state s at time index k=1 to state s¹ , s², s³ at time index k=2. This probability may be termed the success probability P_a(s, s’) and is the probability of going from state s to s’ taking action a. The probability may be determined from the input historical data, particularly the input historical market data. For example, the probability may be calculated as a probability density function from the historical market price and quantity data.

For the UK Balancing Mechanism service, an example source for such historical market data is the Elexon website which includes for all settlement periods all the submitted bids and offers and all the information related to the accepted bids and offers. Details of these bids and offers would include price and the volume which indicates the level of deviation the supplier is agreeing to provide in terms of increasing the generation or decreasing the demand. The probability of success may also be a function of the energy storage characteristics, namely how much energy a particular system could supply for that settlement period. Those energy storage characteristics may be general properties like the set of variables mentioned earlier which are useful for determining bidding strategies, namely min/max power output, min/max cell temperature and cell voltage balance. The degradation model described above predicts the total capacity and associated costs. The determination of the energy storage characteristics could partly be predicted using the degradation model above as well as the nominal values for min/max power output, min/max cell temperature and cell voltage balance.. An estimation of the revenue which will be generated by providing the service is also determined. This may be termed the revenue function R_a(s, s’). The revenue may be generated as a function of the offer or bid price range which is offered to meet the requirements of the identified service as well as the volume offered. The revenue may be obtained from the volume multiplied by the unit price.

Next a set of actions which determine how the battery is controlled over time, i.e. what state the battery should be placed in for participation in the BM service, are obtained. The set of actions may also include a bidding strategy for each time frame, e.g. a price and a volume to be offered. A bidding strategy predicts a suitable bid price and a corresponding success probability. The solution may be based on a Markov Decision Process (MDP) which is a general framework to solve problems in which

• actions taken depend on the system states (e.g. actions = charge/discharge/idle depend on available rewards and also on the future energy storage level which depends on past actions)

• a cumulative reward is optimized (e.g. the reward is a cumulation of the individual rewards at each time index k)

• only the current state (e.g. current battery data) and the past actions are known (e.g. historical data) and

• the system may be non-stationary.

As an example, the Bellman equation may be used:

P_a(s, s’) is the success probability determined above

R_a(s, s’) is the reward function determined above

Y is a discount factor have a value between [0,1] BatDeg is the battery degradation model including cycle counting algorithm with SoC(s’): state of charge as a function of state, T(s’): temperature as a function of state and rC: residual, normalised total capacity of the battery system at the initial state,

Capex: current capex costs for the battery system

EoL: end of life of the battery as determined by manufacturer warranty

BoL: begin of life total of battery system associated with capex value.

The discount factor represents the difference in importance between future and present rewards. In other words, the discount factor values the immediate reward above the future delayed reward and presents the uncertainty about the future. The discount factor is a parameter in MDP which is between [0,1] and is normally set at between 0.8 and 0.9,

A mixed integer linear programming (MILP) algorithm may be used to solve the MDP above. Any suitable algorithm may be used for example as described by VVu, Jianhui, and Edmund H. Durfee in "Mixed-integer linear programming for transition-independent decentralized MDPs." Proceedings of the fifth international joint conference on Autonomous agents and multiagent systems. ACM, 2006.

Figure 9 also shows the profit per settlement period for the energy wholesale market (WSM) which is a relatively flexible short-term market, e.g. bids may be submitted an hour, day, week or month before the settlement period. The revenue from the WSM market may be predicted by standard third-party resources which provide the day ahead market price. For example, this may be calculated using a wholesale market tool comprising an optimization engine which finds the minimum of an objective function F which may be defined as: d d _ 100 x Capex

The SoCd may be determined by selecting the SoC(t) segments where the previous SoC values in the time series is higher and the later SoC values in the time series are lower (i.e. the battery is discharging). Similarly, the SoCc may be determined by selecting the SoC(t) segments where the previous SoC values in the time series is lower and the later SoC values in the time series are higher (i.e. the battery is charging). The BatDeg model may be the one described above. The optimization engine may also include boundary conditions for SoC and battery system availability for each settlement period. Boundary conditions represent any conditions which must be met at the beginning or end of each settlement period. They are discussed in further detail below in relation to the iterative nature of the method.

The result of the optimization is a SoC(t) profile that indicates when and how much power needs to be purchased from or sold to the wholesale market during the next 48 settlement periods in order to maximise profits from this market. Profits are usually realized by a combination of a charge and discharge semi-cycle and therefore the associated profits may be spread over several settlement periods, hence the similar profit over several settlement periods in Figure 9 for the WSM. Negative profits are possible when the associated degradation costs of the battery system are higher than the net trade revenues from buying and selling energy.

Figure 9 also shows the profit per settlement period for the short-term operating reserve (STOR) which is a long-term service, e.g. bids are submitted days or even months before they are implemented. The short term operating reserve (STOR) service (Sam Matthews ; Ivana Kockar, “New Short-Term-Operation-Reserve Services in the UK Electricity Market”, 2007 IEEE Power Engineering Society General Meeting (2007)) requires a bidding strategy, a probability for service delivery (e.g. a success probability that the bid will be accepted) and associated battery degradation costs. The bidding strategy and success probability are obtained from historical data analysis of STOR events in the power grid and from the historical market information which is available, e.g. from the Elexon website, concerning prices and successful bids as a function of time (intraday, week-days/-ends, seasons). The historical data forms the basis for a recommendation system that suggests the bidding price and times and additionally returns a success probability. This historical dataset is continuously updated with new market data for the service to further optimize precision of the recommendation as well as adjust to changes in the service requirements or structure.

The battery degradation model described above is combined with the probability model for a STOR event and the bidding price model to yield the forecast for profits in respect to each settlement period as shown in Figure 9. The combination may comprise using the bidding strategy with the probability of bid acceptance to give a value for the service revenue, e.g. by calculating the revenue value as the probability multiplied by volume multiplied by bid price.

A known service model gives the probability that a certain power needs to be delivered in order to comply with the service (meaning there is a STOR event). The battery degradation model above takes the assumed discharge cycle from the service model to calculate the associated degradation costs. The profit for each time window is then calculated from the sum of the service revenue and the value of the degradation costs multiplied by the probability of a STOR event (this latter value will be negative).

Figure 9 also shows the profit per settlement period for the Firm frequency response (FFR) which is another long-term service. The FFR model for modelling profit may comprise a semi-hidden markov model to anticipate the likelihood of frequency events as described for example by D.M. Greenwood et al. in "Frequency response services designed for energy storage", Applied Energy, Volume 203, 2017, 115-127. Figure 11 is an example of the states within the Markov model and the transition probabilities. There are three types of events: low frequent event, high frequent event and no frequency event. The transition probabilities p are derived from historical data analysis and examples of values are shown in Figure 11. These probabilities may be updated as new data is incorporated into the model. The Markov model yields a probable SoC(t) profile that the battery system is required to comply with in order to avoid penalty charges.

As with the STOR model, from the historical bidding and pricing data, a distribution of successful bids in correlation with timebands (morning, evening, night times, etc...), days (weekdays, weekends) and time of the year (seasonal dependencies of frequency events in the power grid). This data forms the basis for a recommendation system that suggests the bidding price and times and additionally returns a success probability. This historical dataset is continuously updated with new market data from the service to further optimize precision of the recommendation as well as adjust to changes in the service. The battery degradation model is included in the Markov model and combined with the bidding price model to yield the forecast for profits in respect to each settlement period (time window) as shown in Figure 9. The inclusion of the battery degradation model in the Markov model may be similar to that described in relation to the BM service in which the battery degradation costs are part of the Bellman equation, which would be driven from the Markov tree. The optimal result for revenues and degradation costs from the Bellman equation are combined with the bidding price model, which is basically the historical data analysis given the probability that the resulting bids and volume will be successful or not. The probability is then multiplied with the Bellman equation result to give the profit over time graph shown in Figure 9.

Returning to Figure 8, once the individual profit per settlement period has been determined for each of the services and markets, the results from each market and service may be stacked to provide revenue maximisation across the period in question. For each settlement period, the market or service which provides the most profit may be selected. Additionally, any technical constraints which must be met for each market or service together with any legal or service requirements for these markets or services are considered when stacking the individual results. For example, the technical and other legal or service requirements may define boundary conditions between service or market changes.

An example boundary condition is that before switching to supply the STOR service, the battery system is typically required to be at full capacity. Accordingly, if a period of supply to the STOR service is not preceded by a period of charging in the BM service, a short window (e.g. 30 minutes) must be left before the STOR service to allow the charging. The FFR service also typically has upper and lower limits for capacity, e.g. 0.8 or 0.2 and thus a technical constraint before supplying the FFR service is normally to be at a value for the capacity which is mid-way between the two constraints. An example in the BM service is the requirement to give the bids and offers one hour in advance. However, definite information about the status of the battery system (e.g. SoC, cell balance, etc) will not be available. Therefore different (worst case) scenarios need to be considered when submitting bids/offers because not being able to provide the service results in high penalty charges.

The WSM is a short-term market and as described above, the boundary conditions are incorporated into the model for determining the profit. Accordingly, in between long-term services the WSM may be used to comply with boundary conditions like SoC requirements for STOR or FFR services. The BM may be used in a similar way. By contrast to the WSM and partly BM, for the STOR and FFR services, the battery system operator has no influence on the operations after a bid was successful, therefore these services only impose constraints and boundary conditions on the adjacent services, but cannot be used to fulfil any.

Figure 12 shows an example of the actions to be taken (e.g. nature of bids to which markets/services and hence in which markets/services to participate at what time and in what way). For each settlement period (time window), Figure 12 shows the market or service which offers the maximum profit during that time period. For example, for the initial five settlement periods, the optimal service is the FFR service 110 and for the next five settlement periods, the STOR service 112 provides the highest profit. Thereafter, the BM service 116 provides the optimal profit until k=20 when the STOR market is once again optimal. The WSM market is optimal between k=30 and k=40. The structure of the revenue stacking algorithm may be modular so it is possible to remove or add new markets or services without changing the structure of already implemented markets and services.

As shown above in relation to Figure 12, the maximum revenue which can be obtained at each settlement period can be determined for each market or service. Some of the markets and services (e.g. the STOR and FFR services) have legal constraints which require a user to contract to supply the market or participate in the service a long time before the contract must be delivered, e.g. many days or even weeks in advance. Once the user’s offer is accepted, the contract must be delivered to such long-term services at the contracted time period. Other markets and services (e.g. the WSM market and BM service) have shorter time periods, e.g. hours and may thus be termed short-term services. There may be other suppliers who also offer their services and thus not all offers/bids will be accepted. Accordingly, when controlling the battery, it is important to determine which markets and services must be supplied and when as well as how the battery is to be operated when meeting the contract.

Returning to Figure 8, the next step may be to identify the availability of the battery (step S912). This may include determining which of the bids for the long-term services have been successful. Figure 13 shows a graph of battery usage against time with the periods in which the battery must be used to supply a long-term service indicated. As shown in the corresponding graph below, the battery is thus available for supply to short-term services in the gaps between the periods in which the long-term services are being supplied.

When periods of battery availability have been identified, the method may loop back to the start to optimize control of the battery for the remaining time periods. In other words, the steps which lead to obtaining a set of actions S908 may be repeated with the market data obtained in step S902 based on short-term services only. In this loop, the process may be considered to be an optimizer for short-term services.

When no further optimisation is required, e.g. when there is no short-term availability for the battery, the set of actions may be output (step S914). The final output may be a set of actions to control the battery to supply different markets or services in different time periods and to charge or remain idle as appropriate based on the algorithm. The periods when a contract is being provided to different markets or services may be considered to be optimised for revenue.

As explained above, the SoH determined from the degradation model described above may be incorporated into the determinations as to whether to supply a particular market or service. For example, a limited capacity (SoH) which is predicted from the degradation model may impact a decision not to supply or to reduce supply for future markets or services. When incorporating the degradation model, the model may be linearized so that the model is readily combined with other optimisation tools. Linearization of the degradation model can be done using any suitable technique, e.g. the well-known Taylor expansion. The degradation model itself consists of a set of equations that are highly non-linear. The Taylor expansion approximates the equations around a specified value of each of the various variables. If required, quadratic terms in the equation can be linearized by using the McCormick envelope approach, which is a type of convex relaxation technique. For example, if the SoH is expressed as:

SoH = 1 - f cyc_ling (SoC, DoD, Cr, T, Cyc ) - / ca_len_dar (SoC, T, t)

The Taylor expansion gives:

SoHun = 1 — { F(rC ) + a(SoC_cyc ^~ SoC_{0 cyc} ) + b(DoD — DoD₀ ) + c(Cr — Cr₀ ) + d(T — T₀ )

+ e(Cyc - Cyc₀) + f(SoC_cai ^~ SoC_{0 cai} ) + g(.t - t₀)} where the variables and parameters are as defined above.

A comprehensive and accurate comparison of markets/services and their profit potential may be determined by incorporating the market data, the technical constraints and other legal or service requirements of the markets and services as described above,. The optimal use of the battery system may thus be determined. The various requirements may be included via interdependencies of properties such as SoC and opportunity costs of the battery itself. As explained above, the constraints and requirements may define boundary conditions between services or may be incorporated as part of the individual market/service model.

Figure 14 is a schematic block diagram illustrating the components of the system. The system comprises a battery analyser 600 which may perform the methods of Figures 1 or 2 to analyse the degradation of a battery 550 and/or the method of Figure 8 to select how to use the battery. The battery may be an individual battery cell, a battery pack comprising multiple cells or a battery system incorporating multiple battery cells or packs. The battery analyser 600 receives inputs from sensors 500, 502 which measure parameter values for the battery 550. It will be appreciated that the use of one battery and two sensors is merely indicative and the battery analyser may be analysing multiple batteries and receiving information from any number of sensors may be used.

The outputs from the battery analysing, i.e. an indication of the state of the health (SoH) of the battery may be output to a user 700 via any suitable user interface 702, e.g. a screen on a computer or other electronic device. The battery analyser 600 may also be connected to a database 800, which stores for example the training data 820 which is used to train the model as well as the degradation model 814 and the parameters which are most appropriate to be used as a starting set of parameters for a particular battery (e.g. based on manufacturer and chemistry).

The battery analyser 600 may be formed from one or more servers and the steps (or tasks) in Figures 1 and 2 may be split across the one or more servers or the cloud. The battery analyser 600 may include one or more processors 604, one or more memory devices 606 (generically referred to herein as memory 606), one or more input/output ("I/O") interface(s) 608, one or more data ports 610, and data storage 612. The battery analyser 600 may further include one or more buses that functionally couple various components of the battery analyser 600.

The data storage 612 may store one or more operating systems (O/S) 614; and one or more program modules, applications, engines, computer-executable code, scripts, or the like such as, for example, a model engine 616 and a comparison engine 618. The model engine 616 may apply the degradation model and the comparison engine 618 may compare measured and predicted values as described in Figure 1. Any of the components depicted as being stored in data storage 612 may include any combination of software, firmware, and/or hardware. The software and/or firmware may include computer-executable code, instructions, or the like that may be loaded into the memory 606 for execution by one or more of the processor(s) 604 to perform any of the operations described earlier in connection with correspondingly named engines.

The bus(es) may include at least one of a system bus, a memory bus, an address bus, or a message bus, and may permit exchange of information (e.g., data (including computer- executable code), signalling, etc.) between various components of the battery analyser 600. The bus(es) may include, without limitation, a memory bus or a memory controller, a peripheral bus, an accelerated graphics port, and so forth. The bus(es) may be associated with any suitable bus architecture including, without limitation, an Industry Standard Architecture (ISA), a Micro Channel Architecture (MCA), an Enhanced ISA (EISA), a Video Electronics Standards Association (VESA) architecture, an Accelerated Graphics Port (AGP) architecture, a Peripheral Component Interconnects (PCI) architecture, a PCI-Express architecture, a Personal Computer Memory Card International Association (PCMCIA) architecture, a Universal Serial Bus (USB) architecture, and so forth.

The memory 606 of the battery analyser 600 may include volatile memory (memory that maintains its state when supplied with power) such as random access memory (RAM) and/or non-volatile memory (memory that maintains its state even when not supplied with power) such as read-only memory (ROM), flash memory, ferroelectric RAM (FRAM), and so forth. Persistent data storage, as that term is used herein, may include non-volatile memory. In certain example embodiments, volatile memory may enable faster read/write access than non-volatile memory. However, in certain other example embodiments, certain types of non volatile memory (e.g., FRAM) may enable faster read/write access than certain types of volatile memory.

In various implementations, the memory 606 may include multiple different types of memory such as various types of static random access memory (SRAM), various types of dynamic random access memory (DRAM), various types of unalterable ROM, and/or writeable variants of ROM such as electrically erasable programmable read-only memory (EEPROM), flash memory, and so forth. The memory 606 may include main memory as well as various forms of cache memory such as instruction cache(s), data cache(s), translation lookaside buffer(s) (TLBs), and so forth. Further, cache memory such as a data cache may be a multi level cache organized as a hierarchy of one or more cache levels (L1, L2, etc.).

The data storage 612 and/or the database 800 may include removable storage and/or non- removable storage including, but not limited to, magnetic storage, optical disk storage, and/or tape storage. The data storage 612 and/or the database 800 may provide non-volatile storage of computer-executable instructions and other data. The memory 606, the database 800 and the data storage 612, removable and/or non-removable, are examples of computer- readable storage media (CRSM).

The data storage 612 may store computer-executable code, instructions, or the like that may be loadable into the memory 606 and executable by the processor(s) 604 to cause the processor(s) 604 to perform or initiate various operations. The data storage 612 may additionally store data that may be copied to memory 606 for use by the processor(s) 604 during the execution of the computer-executable instructions. Moreover, output data generated as a result of execution of the computer-executable instructions by the processor(s) 604 may be stored initially in memory 606, and may ultimately be copied to data storage 612 for non-volatile storage or into the database 800.

The data storage 612 may further store various types of data utilized by components of the battery analyser 600. Any data stored in the data storage 612 may be loaded into the memory 606 for use by the processor(s) 604 in executing computer-executable code. In addition, any data depicted as being stored in the data storage 612 may potentially be stored in one or more of the datastores and may be accessed and loaded in the memory 606 for use by the processor(s) 604 in executing computer-executable code.

The processor(s) 604 may be configured to access the memory 606 and execute computer- executable instructions loaded therein. For example, the processor(s) 604 may be configured to execute computer-executable instructions of the various program modules, applications, engines, or the like of the system to cause or facilitate various operations to be performed in accordance with one or more embodiments of the disclosure. The processor(s) 604 may include any suitable processing unit capable of accepting data as input, processing the input data in accordance with stored computer-executable instructions, and generating output data. The processor(s) 604 may include any type of suitable processing unit including, but not limited to, a central processing unit, a microprocessor, a Reduced Instruction Set Computer (RISC) microprocessor, a Complex Instruction Set Computer (CISC) microprocessor, a microcontroller, an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), a System-on-a-Chip (SoC), a digital signal processor (DSP), and so forth. Further, the processor(s) 604 may have any suitable microarchitecture design that includes any number of constituent components such as, for example, registers, multiplexers, arithmetic logic units, cache controllers for controlling read/write operations to cache memory, branch predictors, or the like. The microarchitecture design of the processor(s) 604 may be capable of supporting any of a variety of instruction sets.

Referring now to other illustrative components depicted as being stored in the data storage 612, the O/S 614 may be loaded from the data storage 612 into the memory 606 and may provide an interface between other application software executing on the battery analyser 600 and hardware resources of the battery analyser 600. More specifically, the O/S 614 may include a set of computer-executable instructions for managing hardware resources of the system and for providing common services to other application programs (e.g., managing memory allocation among various application programs). In certain example embodiments, the O/S 614 may control execution of one or more of the program modules depicted as being stored in the data storage 612. The O/S 614 may include any operating system now known or which may be developed in the future including, but not limited to, any server operating system, any mainframe operating system, or any other proprietary or non proprietary operating system.

Referring now to other illustrative components of the battery analyser 600, the input/output (I/O) interface(s) 608 may facilitate the receipt of input information by the battery analyser 600 from one or more I/O devices as well as the output of information from the battery analyser 600 to the one or more I/O devices. The I/O devices may include any of a variety of components such as a display or display screen having a touch surface or touchscreen; an audio output device for producing sound, such as a speaker; an audio capture device, such as a microphone; an image and/or video capture device, such as a camera; a haptic unit; and so forth. Any of these components may be integrated into the battery analyser 600 or may be separate. The I/O devices may further include, for example, any number of peripheral devices such as sensors, data storage devices, printing devices, and so forth.

The I/O interface(s) 608 may also include an interface for an external peripheral device connection such as universal serial bus (USB), FireWire, Thunderbolt, Ethernet port or other connection protocol that may connect to one or more networks. The I/O interface(s) 608 may also include a connection to one or more antennas to connect to one or more networks via a wireless local area network (WLAN) (such as W-Fi) radio, Bluetooth, and/or a wireless network radio, such as a radio capable of communication with a wireless communication network such as a Long Term Evolution (LTE) network, WiMAX network, 3G network, etc.

The battery analyser 600 may further include one or more data ports 610 via which the battery analyser 600 may communicate with any of the processing modules. The data ports(s) 610 may enable communication with the sensors 500, 502 and the database 800.

It should be appreciated that the engines and the program modules depicted in the Figures are merely illustrative and not exhaustive and that processing described as being supported by any particular engine or module may alternatively be distributed across multiple engines, modules, or the like, or performed by a different engine, module, or the like. In addition, various program module(s), script(s), plug-in(s), Application Programming Interface(s) (API(s)), or any other suitable computer-executable code hosted locally on the system and/or hosted on other computing device(s) accessible via one or more of the network(s), may be provided to support the provided functionality, and/or additional or alternate functionality. Further, functionality may be modularized differently such that processing described as being supported collectively by the collection of engines or the collection of program modules may be performed by a fewer or greater number of engines or program modules, or functionality described as being supported by any particular engine or module may be supported, at least in part, by another engine or program module. In addition, engines or program modules that support the functionality described herein may form part of one or more applications executable across any number of devices of the system in accordance with any suitable computing model such as, for example, a client-server model, a peer-to-peer model, and so forth. In addition, any of the functionality described as being supported by any of the engines or program modules may be implemented, at least partially, in hardware and/or firmware across any number of devices.

It should further be appreciated that the system may include alternate and/or additional hardware, software, or firmware components beyond those described or depicted without departing from the scope of the disclosure. More particularly, it should be appreciated that software, firmware, or hardware components depicted as forming part of the system are merely illustrative and that some components may not be present or additional components may be provided in various embodiments. While various illustrative engines have been depicted and described as software engines or program modules, it should be appreciated that functionality described as being supported by the engines or modules may be enabled by any combination of hardware, software, and/or firmware. It should further be appreciated that each of the above-mentioned engines or modules may, in various embodiments, represent a logical partitioning of supported functionality. This logical partitioning is depicted for ease of explanation of the functionality and may not be representative of the structure of software, hardware, and/or firmware for implementing the functionality. Accordingly, it should be appreciated that functionality described as being provided by a particular engine or module may, in various embodiments, be provided at least in part by one or more other engines or modules. Further, one or more depicted engines or modules may not be present in certain embodiments, while in other embodiments, additional engines or modules not depicted may be present and may support at least a portion of the described functionality and/or additional functionality. Moreover, while certain engines modules may be depicted or described as sub-engines or sub-modules of another engine or module, in certain embodiments, such engines or modules may be provided as independent engines or modules or as sub-engines or sub-modules of other engines or modules.

The operations described and depicted in the illustrative methods of Figures 1 and 5 may be carried out or performed in any suitable order as desired in various example embodiments of the disclosure. Additionally, in certain example embodiments, at least a portion of the operations may be carried out in parallel. Furthermore, in certain example embodiments, less, more, or different operations than those depicted in Figures 1 and 5 may be performed.

Although specific embodiments of the disclosure have been described, one of ordinary skill in the art will recognize that numerous other modifications and alternative embodiments are within the scope of the disclosure. For example, any of the functionality and/or processing capabilities described with respect to a particular system, system component, device, or device component may be performed by any other system, device, or component. Further, while various illustrative implementations and architectures have been described in accordance with embodiments of the disclosure, one of ordinary skill in the art will appreciate that numerous other modifications to the illustrative implementations and architectures described herein are also within the scope of this disclosure.

Certain aspects of the disclosure are described above with reference to block and flow diagrams of systems, methods, apparatuses, and/or computer program products according to example embodiments. It will be understood that one or more blocks of the block diagrams and flow diagrams, and combinations of blocks in the block diagrams and the flow diagrams, respectively, may be implemented by execution of computer-executable program instructions. Likewise, some blocks of the block diagrams and flow diagrams may not necessarily need to be performed in the order presented, or may not necessarily need to be performed at all, according to some embodiments. Further, additional components and/or operations beyond those depicted in blocks of the block and/or flow diagrams may be present in certain embodiments.

Accordingly, blocks of the block diagrams and flow diagrams support combinations of means for performing the specified functions, combinations of elements or steps for performing the specified functions, and program instruction means for performing the specified functions. It will also be understood that each block of the block diagrams and flow diagrams, and combinations of blocks in the block diagrams and flow diagrams, may be implemented by special-purpose, hardware-based computer systems that perform the specified functions, elements or steps, or combinations of special-purpose hardware and computer instructions.

Program modules, applications, or the like disclosed herein may include one or more software components including, for example, software objects, methods, data structures, or the like. Each such software component may include computer-executable instructions that, responsive to execution, cause at least a portion of the functionality described herein (e.g., one or more operations of the illustrative methods described herein) to be performed.

A software component may be coded in any of a variety of programming languages. An illustrative programming language may be a lower-level programming language such as an assembly language associated with a particular hardware architecture and/or operating system platform. A software component comprising assembly language instructions may require conversion into executable machine code by an assembler prior to execution by the hardware architecture and/or platform.

Another example programming language may be a higher-level programming language that may be portable across multiple architectures. A software component comprising higher- level programming language instructions may require conversion to an intermediate representation by an interpreter or a compiler prior to execution.

Other examples of programming languages include, but are not limited to, a macro language, a shell or command language, a job control language, a script language, a database query or search language, or a report writing language. In one or more example embodiments, a software component comprising instructions in one of the foregoing examples of programming languages may be executed directly by an operating system or other software component without having to be first transformed into another form.

A software component may be stored as a file or other data storage construct. Software components of a similar type or functionally related may be stored together such as, for example, in a particular directory, folder, or library. Software components may be static (e.g., pre-established or fixed) or dynamic (e.g., created or modified at the time of execution). Software components may invoke or be invoked by other software components through any of a wide variety of mechanisms. Invoked or invoking software components may comprise other custom-developed application software, operating system functionality (e.g., device drivers, data storage (e.g., file management) routines, other common routines and services, etc.), or third-party software components (e.g., middleware, encryption, or other security software, database management software, file transfer or other network communication software, mathematical or statistical software, image processing software, and format translation software).

Software components associated with a particular solution or system may reside and be executed on a single platform or may be distributed across multiple platforms. The multiple platforms may be associated with more than one hardware vendor, underlying chip technology, or operating system. Furthermore, software components associated with a particular solution or system may be initially written in one or more programming languages, but may invoke software components written in another programming language.

Computer-executable program instructions may be loaded onto a special-purpose computer or other particular machine, a processor, or other programmable data processing apparatus to produce a particular machine, such that execution of the instructions on the computer, processor, or other programmable data processing apparatus causes one or more functions or operations specified in the flow diagrams to be performed. These computer program instructions may also be stored in a computer-readable storage medium (CRSM) that upon execution may direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement one or more functions or operations specified in the flow diagrams. The computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational elements or steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process.

Additional types of CRSM that may be present in any of the devices described herein may include, but are not limited to, programmable random access memory (PRAM), SRAM, DRAM, RAM, ROM, electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the information and which can be accessed. Combinations of any of the above are also included within the scope of CRSM. Alternatively, computer-readable communication media (CRCM) may include computer-readable instructions, program modules, or other data transmitted within a data signal, such as a carrier wave, or other transmission. However, as used herein, CRSM does not include CRCM.

Although embodiments have been described in language specific to structural features and/or methodological acts, it is to be understood that the disclosure is not necessarily limited to the specific features or acts described. Rather, the specific features and acts are disclosed as illustrative forms of implementing the embodiments. Conditional language, such as, among others, "can," "could," "might," or "may," unless specifically stated otherwise, or otherwise understood within the context as used, is generally intended to convey that certain embodiments could include, while other embodiments do not include, certain features, elements, and/or steps. Thus, such conditional language is not generally intended to imply that features, elements, and/or steps are in any way required for one or more embodiments or that one or more embodiments necessarily include logic for deciding, with or without user input or prompting, whether these features, elements, and/or steps are included or are to be performed in any particular embodiment.

Claims

1. A method for optimising usage of an energy storage system comprising at least one battery, the method comprising: measuring a set of variables for the at least one battery; selecting parameters for a degradation model which predicts degradation of the at least one battery; obtaining a degradation value for the battery using a predicted degradation value which is predicted using the degradation model and the selected parameters; obtaining historical data from one or more services to which the energy storage system is connectable; determining, using the degradation value and the historical data, an optimum state for the at least one battery for each of a plurality of time windows, and controlling the energy storage system based on the determined states for the at least one battery.

2. The method of claim 1 , wherein the optimum state is one of discharge to one of the one or more services, charge to one of the one or more services or remain idle.

3. The method of claim 2, wherein determining the optimum state comprises determining a quantity of energy by which the at least one battery is to be charged or a quantity of energy by which the at least one battery is to be discharged.

4. The method of any one of the preceding claims, wherein determining the optimum state comprises determining any boundary conditions which must be met at a start or end of each time window.

5. The method of any one of the preceding claims, wherein the one or more services comprise at least one short-term service and at least one long-term service and determining the optimum state comprises determining the optimum state for the at least one battery for the at least one long term service for each of the plurality of time windows; determining whether there are any time windows in which the at least one battery is available and is not connected to the at least one long-term service; and when it is determined that there are available time windows, determining the optimum state for the at least one battery for the at least one short-term service in the available time windows.

6. The method of any one of the preceding claims, wherein determining the optimum state comprises determining an optimisation value for each of the one or more services for each of the plurality of time windows.

7. The method of claim 6, comprising selecting the service having the highest optimisation value in each of the time windows.

8. The method of claim 6 or claim 7, comprising determining the optimisation value from a bidding strategy and a probability of success for the bidding strategy.

9. The method of any one of claims 8, comprising determining the bidding strategy and probability of success using a Markov model.

10. The method of any one of claims 6 to 9, wherein the one or more services comprises a balancing mechanism (BM) service and the optimisation value for the BM service is determined by determining a set of actions l(s) from

P_a(s, s’) is the success probability

R_a(s, s’) is the reward function

Y is a discount factor have a value between [0,1],

Capex: current capex costs for the battery system

EoLend of life of the battery as a percentage value [0,100] of the battery as determined by manufacturer warranty BoL: begin of life as a percentage value [0,100] of the battery system associated with capex value.

11. The method of any one of claims 6 to 10, wherein the one or more services comprises a wholesale market (WSM) service and the optimisation value for the WSM service is determined by minimising an objective function F defined as: d d _ 100 x Capex

BatDeg: Battery degradation model including cycle counting algorithm with SoC(t): state of charge as a function of time T(t): temperature as a function of time rC: residual, normalised total capacity of the battery system at the begin of the day Capex: current capex costs for the battery system

BoL: begin of life as a percentage value [0,100] of the battery system associated with capex value

12. The method of any one of the preceding claims wherein the degradation model comprises a calendar ageing component and a cycling ageing component and obtaining the degradation value comprises obtaining an estimated degradation value for the battery using the set of measured parameters; and outputting a degradation value based on the estimated and predicted degradation values.

13. The method of any one of the preceding claims, wherein the measured variables comprise at least one of current, voltage, state of charge, depth of discharge, temperature, number of cycles, CP-rate, minimum power out, maximum power out, maximum temperature, minimum temperature, maximum cell voltage balance, minimum and maximum SoC..

14. A computer readable medium carrying processor control code which when implemented in a system causes the system to carry out the method of any one of claims 1 to 13.

15. A battery optimisation system comprising at least one sensor for measuring a battery parameter; a processor which is configured to carry out the method of any one of claims 1 to 13, and a user interface which is configured to display the output result which is generated by the processor.