WO2013104765A2

WO2013104765A2 - Device profile optimization device and method

Info

Publication number: WO2013104765A2
Application number: PCT/EP2013/050503
Authority: WO
Inventors: Chris Minnoy; Bert Robben; Geert PREMEREUR; Michel TILMAN; Alfred Spiessens
Original assignee: Sony Corporation; Sony Europe Limited
Priority date: 2012-01-13
Filing date: 2013-01-11
Publication date: 2013-07-18
Also published as: WO2013104765A3

Abstract

The present invention relates to a device profile optimization device (60) and a corresponding method for optimizing device profiles of one or more electrical devices (20, 22, 24), each device profile including energy usage-related and/or energy output-related information of an electrical device (20), comprising a measurement data input (61) that obtains measurement data of electrical parameters of an electrical device (20) to be optimized and/or of installations electrically coupled to said electrical device, and an optimization unit (62) that optimizes the device profile by use of said measurement data into an optimized device profile.

Description

DEVICE PROFILE OPTIMIZATION DEVICE AND METHOD

FIELD OF THE INVENTION

[0001] The present invention relates to a device profile optimization device and method for optimizing device profiles of one or more electrical devices, each device profile including energy usage-related and/or energy output-related information of an electrical device. The present invention relates further to a control system and a control method. The present invention relates further to a computer program and a computer readable non- transitory medium.

BACKGROUND OF THE INVENTION [0002] Until now, demand-response has been mostly provided by out-of-band communication. For example, a utility company, often through intermediaries, notifies a large number of smaller customers that they should manually turn off electrical devices agreed upon ahead of the demand-response event. A step up is where such devices are attached to a communication network and are triggered by messages sent over said communication network. It is possible to use the electricity network itself as the communication network. The known approaches are mostly useful for reducing extreme peaks on the electrical power distribution network.

[0003] Another approach is where devices negotiate with a party whether and how much electrical energy they should consume or make available. Past and current measurementscan be taken into account to reduce both consumption beyond and below previously anticipated levels. The most advanced systems integrate predictions of external factors such as temperature, wind speed and cloud coverage together with general and averaged estimates of power consumption and generation. Besides reducing deviations against provisioned power levels, the more advanced systems can also be employed to target other goals and/or a different target audiences, for example minimising the amount a large energy consumer has to pay for the energy it buys from a provider.

[0004] US 2011/0060476 A 1 discloses an energy management system which calculates an adjustment amount of energy in a case where an energy supply device requests an energy demand device to adjust the amount of energy to be supplied. Data available in a so-called score card concerning an adjustable energy amount may be used for negotiation to regulate the energy adjustment amount.

[0005] In general, the quality of a demand response approach is heavily dependent on the quality of the available data describing the energy usage-related and/or energy output-related behaviour of electrical devices taking part in the demand response scheme. Such data may be included in device profiles a control device has access to. Those device profiles should match the actual characteristics of the electrical devices as closely as possible. Generally, however, device profiles do not reflect differences between individual electric devices of the same brand and type, which may e.g. result from usage history, aging and environmental factors.

BRIEF SUMMARY OF THE INVENTION

[0006] It is an object of the present invention to provide an improved device profile optimization device and a corresponding method that enable an effective optimization of device profiles. It is a further object of the present invention to provide a corresponding control system and method.

[0007] According to an aspect of the present invention there is provided a device profile optimization device comprising:

a measurement data input that obtains measurement data of electrical parameters of an electrical device to be optimized and/or of installations electrically coupled to said electrical device, and

an optimization unit that optimizes the device profile by use of said measurement data into an optimized device profile.

[0008] According to a further aspect of the present invention there is provided a control system comprising:

a control device that controls one or more electrical devices,

one or more measurement data sensors that measure measurement data of electrical parameters of said one or more electrical devices and/or installations electrically coupled to said one or more electrical devices, and

a storage unit that stores device profiles of said one or more electrical devices, a device profile optimization device as proposed according to the present invention for optimizing the device profile of one or more of said one or more electrical devices, each device profile including energy usage-related and/or energy output-related information of an electrical device, wherein said control device is configured to control an electrical device by use of the corresponding optimized device profile.

[0009] According to further aspects a corresponding device profile optimization method and a corresponding control method are provided according to the present invention.

[0010] According to still further aspects a computer program comprising program means for causing a computer to carry out the steps of the method according to the present invention, when said computer program is carried out on a computer, as well as a computer readable non-transitory medium having instructions stored thereon which, when carried out on a computer, cause the computer to perform the steps of the device profile optimization method according to the present invention are provided.

[0011] Preferred embodiments of the invention are defined in the dependent claims. It shall be understood that the claimed control methods have similar and/or identical preferred embodiments as the claimed device profile optimization device and as defined in the dependent claims.

[0012] The invention is applied in a demand response environment where one or more (electricity consuming and/or producing) electrical devices from one or more different vendors and/or brands are controlled by means of a controlling party (i.e. a control device) external to the electrical devices. The controlling party may use some prediction of the future to optimize e.g. the timing of power and/or energy allocations to be assignedto the electrical devices taking part in a demand-response scheme. In such an environment, the controlling party (i.e. the control device) preferably has access to optimized device profiles to perform the control in a near-optimal fashion. Generally, in the same way, by the same control device and/or by use of the same (optimized) device profile one or more electrical devices can be controlled. Preferably, such optimized device profiles are provided by a device profile optimization device according to the invention. For example, the device profile optimization device according to the invention may enhance (more) generic device profiles so that they are adapted to one or more specific devices in a specific environment and are thus optimized. In this context, the term "optimize" shall not be understood in the sense that necessarily the absolute "optimum" is reached by the "optimization", but shall be understood in the sense that an improvement or enhancement of a previous solution is found. For example, device profiles may be repeatedly enhanced ("optimized") , wherein each repetition may refine and improve the device profiles with regard to the one or more electrical devices it is associated with, e.g. by taking into account measurement data reflecting the specific conditions under which the one or more devices operate.

[0013] Particularly the device operation, the electricity consumption and/or production of the electrical devices may be controlled, e.g. optimized, in view of demand response considerations and based on the optimized device profiles. For the control separate control models may be used, e.g. a different model for each different kind of demand response goal to be achieved. Such demand response goals might e.g. be minimizing local energy cost or avoiding blackouts. The optimized device profiles and the control may also take into account the surroundings of an electrical device to be controlled, the climate of the surroundings, and other factors related to the device.

[0014] According to the present invention measurement data of electrical parameters of said electrical device are obtained (i.e. received or fetched) via a measurement data input (e.g. a data interface) and used for optimizing a device profile. In particular, measurement data of electrical parameters of said electrical device and/or of one or more installations being electrically coupled to said electrical device are used. Such installations may e.g. be other electrical devices in the vicinity of said electrical device (e.g. within a certain distance from said electrical device or located in the same building, apartment, street, or belonging to the same household as said electrical device) or the electricity distribution center within an apartment, a building or a factory or the buildings within a complete street or village including said electrical device. [0015] Preferably, environmental data about the environment of said one or more electrical devices are used in addition. In this context environmental data about the environment of said electrical device shall be understood as any data about physical conditions in the vicinity of the electrical device to be optimized that may have an influence on the operation of said device and that may be relevant to be taken into account for the optimization of the respective device profile. For instance, the environmental data may include one or more of temperature, atmospheric pressure, humidity, date, time, location and weather conditions.

[0016] Also, data about the state of the electrical device, such as load factor of an electric boiler or selected program of a washing machine, are preferably included in an optimized device profile.

[0017] Moreover, feedback about responses to applied device stimuli, including context data from line properties, voltage, line voltage frequency (spectrum), and other, external factors that may influence the measurements, may be collected and later processed and (e.g. statistically) analyzed in order to derive meaningful cause-effect relations and, thus, to optimize the device profile of one or more electrical devices.

[0018] In this context a device profile shall be understood as a collection of device-related information, particularly including energy usage-related and/or energy output- related information of an electrical device. In particular, energy usage-related and energy output-related information may be information which characterizes the (e.g. past and/or current and/or expected) energy (or electricity or power) consumption and/or the production of an electrical device, e.g. the power consumption and/or production over time. In addition, the device profile may indicate flexibilities as regards the time and/or amount of energy or power consumption and/or production. Furthermore, the device profile could include e.g. technical or user-specific constraints. [0019] A device profile may thus, generally, include one or more of an attribute value concerning the adjustable energy amount, for example information of an adjustable amount of an energy consumption (utilization) of the device, information of a shiftable amount in a time direction in a case where thetime period when the device consumes electricitycan be shifted, information of an operation time of the device (the utilization time of the energy) and/or an adjustment amount of an accumulated energy of the device, information of a propriety (possibility) of blocking of the energy consumption of the device (forced load blocking), information of the consumption of the deviceor an amount of electricityto be generated, and/or an error amount or an error ratio thereof (a specific value of a variance range of the energy consumption of a devicewhose demand cannot be predicted such as a non-networking device or a variance range of the amount of the power to be generated by a distributed power source whose amount of the power to be generated cannot be predicted such as solar power generation or the like). Generally spoken, the device characteristics that let an optimizer (control device) come up with an optimized prosumption plan may be included in an optimized device profile. The optimized device profiles may also reflect the characteristic energy consumption (or variations thereof) of the one or more devices, variations in the manufacturing / operation of the devices, other constraints of the devices, any known upcoming tasks for the devices, the control, the technical infrastructure and/or the user preferences.

[0020] Preferably, a device profile is not a fixed set of power production/consumption attribute values over time, but a recipe for the way an electrical device behaves given certain control inputs. The more flexibility, i.e. the greater and the more fine-grained variations, the device profile allows, the more valuable the device profile is. A device profile can be expressed in multiple ways. For example by functions in the mathematical sense that describe the device's capabilities in full detail or by enumeration of several sets of possible energy states discretized over time. Profiles can include a preference for one or more or all of the various operating schedules they describe. [0021] Optimization of device profiles, as e.g. provided by the device profile optimization device, may e.g. replace the (or some of the) current attribute values of a device profile by other values that allow an optimizer (control device) to take into account a more correct prediction of the electrical future behavior of an electrical device, e.g. in case said control device should apply a certain stimulus to said electrical device.

[0022] In general, types of attribute values that reflect e.g. available prosumption flexibility for electrical devices that possess demand-response flexibility are very valuable for planning the operation of said electrical devices. Such attributes values may be obtained by measuring electrical and environmental properties and enhancing (optimizing) device profiles correspondingly. However, different kinds of attribute values may be obtained by different means and used for enhancing (optimizing) device profiles; the invention is not limited to just enhancing attribute values based on knowledge about electrical and environmental information obtainable by the above measurements.

[0023] For instance, in the case of a washing machine, the device profile of the washing machine may include, per program and sub-program, the sequence of subprograms, the duration, the energy usage and/or energy output requirements and, to indicate flexibilities, the minimum and maximum allowed times until the next sub-program is run. Different wait times may be assigned different priority values to express preference. Another example concerns charging batteries. A device profile for a battery could indicate flexibilities, since batteries may be charged normally, but they may also be charged using fast charging. Further, while batteries may be charged in one go or in several sessions, constraints on the minimal uninterrupted and maximal continuous charging time in any charging session may be imposed. Still another example is the device profile of an electric car to be charged at home overnight. The device profile, for instance, depends on the battery characteristics and state. A typical car battery with 50% charge level may take up to 4 hours of normal charging. The proposed control device is allowed to plan the charging session in such a way that some goal, e.g. lowest charge cost, is reached (e.g. in 16 blocks of 15 minutes each with pauses in between instead of a single, continuous block of 4 hours; the single block option might have a higher preference however as it causes less wear on the components of the car battery). Constraints in the device profile may impose limits for start and end time or, as a technical constraint, a minimum energy amount needed to enable proper charging. In some cases, additional constraints are provided external to the device profile (e.g. the maximum amount of power that can be sustained by the electrical wiring supplying power to the electrical device).

[0024] Thus, generally a device profile captures the expected energy consumption and production characteristics of a device, and may include intrinsic, device-related constraints. Device profiles may assume a closed-world view, whereby each device is independent from other devices, but properly configured for their environments. Thus, default profiles may be used as well, at least initially, which may then be further improved, e.g. by means of the device profile optimization device. A simple default device profile may be a block of constant power production/consumption with a value for the average power production/consumption during some time period. In other embodiments, (optimized) device profiles reflect the surroundings and/or environment of their associated devices, e.g. other devices connected in their vicinity, in the same home network or similar. In this case, the profiles still describe the behavior and flexibility of a single device, but take into account how said behavior and flexibility are influenced by the environment in which the device operates. For example, the same type of washing machine during the winter in a Nordic country will be supplied with colder water and thus require more electric energy to heat water than one in an equatorial country. In addition, the device profile may include different kinds of operating parameters of the device.

[0025] A device profile for an electrical device can also cater for variations in device prosumption based on the properties of the upcoming task for said device. Taking a tumble dryer as an example, not only the properties of a chosen program may be reflected in an optimized device profile, but also, as a refinement and when measurements are available, the effects of the mass and humidity of the clothes placed in the drum of the tumble dryer. [0026] The operation of the control device, which may be part of the proposed control system or at least connected to the control system, i.e. how the control commands are generated from a device profile of an electrical device, is generally known in the art and is, for instance, described in ISO/IEC 14543-3 ( NX) or ISO/IEC DIS 14908 (LON).

[0027] In this context it shall be noted that the expressions "control device" and "control system" shall not be understood in the sense that an electrical device is completely or directly "controlled" by the control device. While this is generally possible, the control may also take place indirectly. For example, an electrical device has its own built-in control device for general operation and control of the electrical device. In this case, the control device may provide the control commands to the built-in control device that will then govern the way in which it operates and controls the electrical device based on these control commands. In addition, an "optimization" of an electrical device controlled by a control device shall be understood in this context as using said electrical device in such a way that it contributes optimally, within the limits of a certain algorithm, to a goal to be achieved by said control device.

[0028] The control commands may e.g. comprise power prosumption (prosumption = production and/or consumption) functions for the electrical devices that are used to optimize the quantity and scheduling of the energy prosumption of the devices. They may be represented as discrete lists of power values (e.g. expected power consumption for every 5 minutes), or as actual mappings from the time domain to power values, in a form suited to the algorithms (e.g. linear programming) used by the control device. In other embodiments the control commands may be on a higher level, e.g. start/stop/pause/resume, start a certain program, operate at certain rotations per minute, charge at a certain voltage / current, etc.

BRIEF DESCRIPTION OF THE DRAWINGS [0029] These and other aspects of the present invention will be apparent from and explained in more detail below with reference to the embodiments described hereinafter. In the following drawings

Fig. 1 shows a schematic diagram of a first embodiment of a control system according to the present invention,

Fig. 2 shows a schematic diagram of a first embodiment of a device profile optimization device according to the present invention,

Fig. 3 shows a schematic diagram of a second embodiment of a control system according to the present invention,

Fig. 4 shows diagrams of the power usage over time of different electrical devices,

Fig. 5 shows a diagram of the power usage over time of a combination of electrical devices,

Fig. 6 shows a diagram of the power usage over time of a washing machine as derived from a number of measurements versus a ideal diagram,

Fig. 7 shows two flow charts illustrating the control methods according to the present invention, and

Fig. 8 shows a schematic diagram of a third embodiment of a control system according to the present invention.

DETAILED DESCRIPTION OF THE INVENTION [0030] The present invention is applied in a demand response environment where electrical devices are controlled by means of a party external to the device. One example of such an environment is a home energy network that combines electrical appliances, measurement devices and a control device, e.g. implemented as a software program on a controller, processor or computer or implemented as dedicated hardware such as an integrated circuit, which can control the electrical devices, in particular influence when appliances are turned on and off, if and when energy usage-related information and/or energy output-related information, e.g. operating parameters, are changed, etc. More specifically, the present invention preferably operates in a setting where a controlling party (i.e. the control device) uses some (expected) knowledge of the future to optimize the timing of power allocations sent to electrical devices taking part in the demand-response scheme. The more accurate and precise the information is a controlling party has about the future, the better it can optimize how and when devices are controlled and used. For example, such information (or, to be more precise, prediction) of the future may generally specify the effects of sending control commands to electrical devices, which is covered by device profiles.

[0031] It is quite typical for many electrical devices that the energy they consume once they enter a certain state is rather predictable, since it is tied to the physical process they implement. Also, the same device can exhibit different usage patterns depending on the mode of use. For instance, considering as an example a modern washing machine, it has many different programs, each of which consists of sub-programs that use different amounts of energy for different lengths of time. Generally, there is no time between these sub-programs for most washing machines since they try to minimize the total duration. However, in a monetary or network optimization strategy it can be useful to also let the time gap between sub-programs vary.

[0032] As mentioned above, all the device characteristics that let an optimizer (i.e. control device) come up with an optimized solution (e.g. an energy consumption and production schedule optimized in view of demand response considerations) can be grouped in a device profile. In the case of the washing machine example this is, per program and sub-program, the sequence of sub-programs, the duration, the energy usage requirements and the minimum and maximum allowed times until the next sub-program. The device characteristics may partly vary for different types of electrical machines, e.g. washing machines, dish washers, TV sets, cooking appliances, refrigerators as used in private households or all kinds of professional machines like robots or electrical motors in factories.

[0033] Also, energy-outputting devices such as batteries and combined heat power devices (CHP, also called cogeneration devices) have their own characteristics with regards to the amount, duration and preference with which they can output electric energy. In case of a battery, a single device can even both take in and output electric energy (or power). A battery device profile describes e.g. under what circumstances the battery can load or unload, what the minimal and maximal power is over time, what the minimal and maximal capacities are, how long and short a load/unload cycle should be etc.

[0034] Fig. 1 shows a schematic diagram of a first embodiment of a control system 10 according to the present invention. Generally, the control system 1 comprises a control device 10 that controls one or more electrical devices 20, 22, 24. Further, the control system 1 comprises one or more measurement data sensors 30, 32, 34 that measure measurement data of electrical parameters of said one or more electrical devices 20, 22, 24 and/or installations 40, 42 (in this example a building 40 and a power distribution device 42) electrically coupled to said one or more electrical devices 20, 22, 24. Further, a storage unit 50 (also called profile repository) is provided that stores device profiles of said one or more electrical devices 20, 22, 24. Finally, a device profile optimization device 60 is provided for optimizing the device profile of one or more of said one or more electrical devices 20, 22, 24.Adevice profile includes energy usage-related and/or energy output- related information of an electrical device. Said control device 10 is configured to control an electrical device by use of the corresponding optimized device profile. [0035] The measurement data sensors 30, 32, 34 e.g. continuously or periodically monitor an electrical device 20 to be controlled by control device lOor measure measurement data of electrical parameters of an electrical device 20 to be controlled e.g. continuously, periodically or when said electrical device 20 operates.

[0036] Fig. 2 shows a schematic diagram of a first embodiment of a device profile optimization device 60 according to the present invention. It comprises a measurement data input 61 that obtains measurement data of electrical parameters of said electrical device and/or of installations electrically coupled to said electrical device and an optimization unit 62 that optimizes the obtained device profile by use of said measurement data into an optimized device profile.

[0037] The invention addresses the problem that in order to most effectively control electricity usage in an installation, such as a building 40 or an industrial site, as much as possible should be known about the behavior of the electrical devices 20, 22, 24 within that installation over time. This information is captured and used to optimize the device profiles of said electrical devices 20, 22, 24.

[0038] One fixed device profile for all electrical devices of a same type (e.g. provided by a vendor) is an acceptable base line, but in many cases significant improvements are possible. For instance, in case of a washing machine with many different programs it is required to at least differentiate for the actual task of the washing machine. It would be even better if device profiles were tailored to individual appliances, not so much because individual electrical devices of the same model using identical operating schedules exhibit different behavior, but because of the different environments in which they are operating. For instance, if two washing machines A and B are considered and if the typical water inlet temperature of washing machine A is 10° C higher than that of washing machine B, then this will lead to a significant power usage reduction in the water heating phase of identical washing programs. [0039] According to the present invention, the effect of initiating a certain command to an electrical device on the local electrical network can be measured, preferably grid-wise close to the electrical device that is controlled. Initially, a pre-defined device profile may be used to schedule the actions of the controlled electrical device, but at the same time the effects of those actions are measured. This information (as a kind of feedback) is used for optimizations of subsequent scheduling runs of the same electrical device.

[0040] Fig. 3 shows a schematic diagram of a second embodiment of a control system 2 according to the present invention. The control system 2 substantially comprises the same elements as the control system 1 shown in Fig. 1, but additionally comprises further elements. In particular, one or more environmental data sensors 70, 72 are provided that measure environmental data about the environment of said one or more electrical devices20, 22, 24 which are input to the profile optimization unit 60 via an environmental data input 6Γ.

[0041] The electrical devices 20, 22, 24 and the measurement data sensors (gauges) 30, 32, 34, 36 are part of a building 40. The measurement data sensors are from left to right farther away from the electrical device 20 whose effect they are measuring and whose device profile - as an example - shall be optimized. A first measurement data sensor 30 is arranged within a first electrical device 20. A second measurement data sensor 32 is arranged in close proximity, e.g. in the power line 33, of a second electrical device 22. A third measurement data sensor 34 is attached to a whole circuit within the building 40, e.g. in the power line 35 within the building 40. A fourth measurement data sensor 36 is arranged in the mains entering the distribution cabinet 42. The building 40 and the distribution cabinet 42 represent installations. Alternatively, the devices 22, 24 and/or the measurement data sensors may represent installations.

[0042] In this embodiment the control device 10 (also called local optimizer), which is preferably arranged within the building, comprises a control unit 12 (also called device steering unit) that controls said electrical devices 20, 22, 24, a measurement data collection unit 14 that obtains the measurement data (e.g. electrical voltage, electrical current and line voltage frequency of one or more electrical devices, a group of electrical devices, a building, a group of buildings including said electrical device, and/or an electrical network to which said electrical device is connected) from said measurement data sensors 30, 32, 34, 36 and an environment data collection unit 16 that obtains the environmental data (e.g. temperature, atmospheric pressure, humidity, date, time, location and/or weather conditions) from said environmental data sensors 70, 72 arranged within and/or outside said building 40.

[0043] Further, one or more (optional) device sensors 80 are provided in this embodiment with the electrical device 20 that measure load characteristics of the electrical device. For instance, in case of a tumble dryer, a device sensor 80 may be provided that measures humidity and mass of the clothes put into the tumble dryer for the next tumble operation. In another embodiment the may be several measurement data sensors (like the sensor 30) within the electrical device 20, and one or more of such sensors may be able to fulfil the task of the device sensor(s) 80.

[0044] The different data can be obtained through dedicated wiring (most likely when there is only one or a few gauges inside a building 40, e.g. one in the cabinet), over power line or over a wired or wireless IP connection (in case of smart devices), as indicated by the arrows pointing from the gauges 30, 32, 34, 36, the environmental data sensors 70, 72 and the device sensor(s) 80 to the respective data collection unit 14, 16 in the control device. It shall, however, be noted that these data collection units 14, 16 may also be part of the device profile optimization device 60. The control unit 12 may or may not use the same communication protocol, but is in any case directed to the individual electrical devices 20, 22, 24, as indicated by the different arrows emanating from the device control unit 12

[0045] The control device 10 makes use of the services of a profile repository 50, serving as storage unit, for downloading appropriate device profiles the control device 10 uses for scheduling its steering actions. The profile optimization device 60 may have a link to and may communicate with the control device 10 and thus knows when an electrical device has been told to start/stop. This may help in determining what part of the power draw on a circuit can be attributed to what electrical device because there is a an evident link between cause and effect. Otherwise, it is virtually impossible to determine exact timings. For example, it may be supposed that a start command is sent to a device, but that the device needs to do some internal setup before it starts 2 minutes later. Without the timing of the initial command there is no evident way of knowing that the device starts with a usage of 0W for 2 minutes. If this information is not captured, it cannot be used in later optimisation steps and all usage will be 2 minutes off.

[0046] The control device 10 sends the obtained measurement data and the obtained environmental data, preferably together with measurement meta-data (e.g. the time of measurement, its accuracy, its sampling frequency, etc.), to the device profile optimization device 60 (also called profile refiner). In the embodiment of optimization device 60shown in Fig. 3 it validates and pre-processes the obtained measurements (e.g. unit conversion) in a measurement preprocessor 63. Further, tags may be added by said measurement preprocessor 63 to the measurement data, said tags including e.g. a distance identifier indicating the distance from said electrical device 20 at which the respective measurement data are measured.

[0047] Generally, any contextual data can be added to the measurement data as tags. It is up to the statistical algoritm to determine whether there is a sufficient correlation between the presence and/or value of a tag and the variations in the measurement data. This makes it possible that a later application of the device profile yields data that is dependent on the particular value of a tag. For instance, in case of a washing machine it may be measured what power it draws many times. Every time this data may be logged and tagged with the inlet water temperature. Over time, a dependency between the inlet water temperature and the power draw might be established (e.g.it might be noticed that the lower inlet water temperature corresponds with longer maximal power use for heating the, colder, water). In general,there is not a fixed set of tags that would limit possible correlations and devices. It is clear that when a profile is subsequently used in an optimisation session, better results can be achieved by supplying a set of values for tag properties that describe the circumstances in which the profile is to be applied. These tags allow retrieving the best matching profile.

[0048] The (preprocessed and/or the unprocessed) measurements are stored in a measurement storage unit 64. When enough new measurements are available, an analysis phase is initiated in a measurement analysis unit 65 preferably in said optimization unit 62 shown in Fig. 2. Herein, a profile fetcher unit 66 fetches one or more device profiles that likely apply to the measurements from the profile repository 50 for use as reference data (e.g. because the measurement data comes from sensors in the vicinity of the corresponding devices etc.). Then the measurements are pooled and subjected to an appropriate statistical algorithm in the measurement analysis unit 65. Preferably, additional energy usage-related and/or energy output-related information (e.g. power consumption over time, voltage over time, current over time, etc.) about the operation of the electrical device 20 is derived. From the measurement data, the environmental data and (optionally derived) energy usage-related and/or energy output-related information a new (optimized) set (the set can contain one element) of device profiles together with a measure of correctness is obtained and clustered around the tags sent by the control device 10. This measure of correctness (sometimes also called "probability measure") needs not be a single measure. For instance, one could be quite certain about power usage in the first half of a device profile and less in the second half the same device profile.

[0049] If there are many measurements for the same device, the measurements can be divided in groups or clusters. The grouping should be done on similarity of the measurement data. It may then become clear that the measurements in the groups have values for the tags that are also similar within the group, but different compared to other groups. If this is the case, it can be said that measurements are clustered around a tag. [0050] Finally, a profile uploader 67 sends the results of the analysis that attain a minimal quality level back to the profile repository 50 that can either accept them and replace previous device profiles or reject them as it sees fit.

[0051] The connections between the control device 10 and the various sensors 30, 32, 34, 36, 70, 72 and 80 may enable wired or wireless communication between the respective components. Preferably, the latency on the connections should not be too high. Preferably, the optimizer, i.e. the control device 10, should have a low-latency connection to the electrical devices and should have as much knowledge as possible about a device in order to effectively manage it. However, to make the optimizer as universally applicable as possible, it should only rely on standardized data and not require device-specific rules. That data is captured in a device profile. A plurality of devices is preferably controlled by a local optimizer, e.g. in the same room, house, building, factory, city, etc. using individual device profiles. In general, however, the location of the optimizer is not relevant. The location of any sensors, e.g. for sensing any parameters or conditions (see below), is generally more relevant. For example, measuring voltage at the home entrance will have less accuracy in determining the patterns of a dishwasher than when measuring the voltage at the plug of the dishwasher.

[0052] Generally, a device profile is used to optimize electricity production/consumption. The device profile may describe the constraints and cause/effect relationships between receiving a command and the resulting power draw/generation. Generally, it does not specify when a certain command should be sent. That is the role of the local optimizer that considers e.g. a number of devices, their individual profiles, the target consumption/generation and optimization algorithm parameters. In an example a washing machine profile shall be considered, said washing machine profile specifying that a certain program starts by using 500W for 3 minutes followed by 2000W for 12 minutes. There are two identical washing machines available to control. An optimizer having a target that 350Wh should be consumed in the first quarter of an hour will start the first washing machine immediately (500W*3/60h+1200W*12/60h=265Wh), while the second washing machine is started after 9 minutes (500W*3/60h+1200W*3/60h=85Wh) for a total of 350Wh. This scenario shall be compared with an optimizer that has no profile information, but can only measure the effect of starting a washing machine. It starts a washing machine immediately (again 265 Whin 15 minutes) and another washing machine immediately after measuring the effect of starting the first washing machine (just a bit less than 225Wh in 15 minutes). It does so because, having no view of the future through the profile, it expects that the first washing machine will keep on using 500W for the next 15 minutes. The result is a usage of 530Wh instead of the intended 350Wh. Algorithms for finding such an optimal solution are widely known and can be based on "constraint programming" as e.g. described in Paul Shaw "Using Constraint Programming and Local Search Methods to Solve Vehicle Routing Problem", CP '98 Proceedings of the 4th International Conference on Principles and Practice of Constraint Programming, Springer Verlag, 1998. If many devices are involved approximations to the optimal solutions have to be searched. One technique is solving a multi-dimensional bin packing problem e.g. as described in Chandra Chekuri and Sanjeev Khanna "On Multi-dimensional Packing Problems", SIAM Journal on Computing, Volume 33, Issue 4, 2004, pages 837-851.

[0053] In an embodiment of the present invention feedback about responses to applied device stimuli is recorded and used to improve (optimize) the device profile for a subsequent application. This feedback can be applied to a subsequent run of the same device instance or other similar devices on or off the local network. In this context "stimuli"mean the commands the control device sends in its normal course of operation using device profiles which are currently active.

[0054] Given that electrical devices typically share the electricity network with many other electrical devices, the effect of applying an external stimulus to an electrical device can often not be measured in isolation. Therefore, in an embodiment of the invention the line properties, voltage and line voltage frequency, are measured as is and later processed by an profile refiner that e.g. does statistical analysis on the measurement data in order to derive meaningful cause-effect relations for use in the improvement of device profiles.

[0055] Further, since external factors, such as temperature, atmospheric pressure, humidity etc., may influence the physical processes that are executed by the electrical device to be optimized, these variables are measured as well by use of environmental sensors.

[0056] Finally, the measurement data are preferably tagged with an identifier for the selected device profile and/or program of operation of the respective electrical device.

[0057] The measurement data are preferably collected by the electrical device itself or as close to it as possible.Preferably,the measurement data is tagged with an indication of how near to the electrical device it was collected. The tags are generally used to add another dimension to the data. For instance, the power measurement may beconsidered as the first dimension, the temperature as the second dimension, the time as the third dimension etc. Generally, such a tag or its dimension, respectively, needs not to be known beforehand.

[0058] After the measurement data is collected, it is handled by a profile optimization device 60 that e.g. uses the statistical processing mentioned before on many captured measurement data sets to derive actual device characteristics. This new (analysed) data is combined in a new version of the device profile and preferably fed back to the profile repository and/or used by the control device 10 to control the respective electrical device. The control device 10 generally processes an obtained device profile and generates control commands for controlling the respective electrical devices based on the obtained device profile. Alternatively, a control device may also derive from the device profile that it is not opportune to send control commands at a certain time. [0059] As shown in Fig. 3, an embodiment of the profile optimization device 60 comprises a measurement analysis unit 62 which preferably performs a statistical analysis. This analysis takes in noisy measurement data and extracts ready-to-use device profiles. There is a whole field of data-analysis and data-mining dedicated to the problem of extracting meaningful information out of noisy measurement data so that many implementations of the analysis can be used in such embodiments, the present invention being not limited to any particular implementation. One usable implementation is e.g. described in "Disaggregated End-Use Energy Sensing for the Smart Grid [IEEE Pervasive Computing, Special Issue on Smart Energy Systems, to appear in Jan-Mar 2011 issue, Jon Froehlichl, Eric Larson, Sidhant Gupta, Gabe Cohn, Matthew S. Reynolds, Shwetak N. Patel]". In the following a simple algorithm (ensemble averaging) shall be explained to clarify and to show that device profiles can be extracted even for very confounded data.

[0060] The ideal power usage of a certain program of a washing machine, as depicted over time in Fig. 4A, shall be used as starting point. Power usage is a very important part of a device profile. Further, the device profiles of a refrigerator (as shown in Fig. 4B), an electric oven (as shown in Fig. 4C) and a radio (as shown in Fig. 4D) shall be considered. If the combination of all four devices (with the refrigerator, radio and electric oven randomly shifted) is measured the diagram of the power usage shown in Fig. 5 is obtained.

[0061] The actual washing machine profile is quite obscured in this combined power usage profile shown in Fig. 5. But if a number (e.g. 10) of such measurements are taken, to which a very simple algorithm, such as using the minimal value of all measurements, is applied the processed ("derived") profile shown in Fig. 6 versus the ideal profile is achieved. As can be seen, the ideal and derived profiles are almost identical. Hence, it is demonstrated that without using any a priori knowledge about the profile of the electrical device that is actually measured, a very close approximation to the actual profile can be found. The knowledge that was used is the starting time of the electrical device. Furthermore, it was assumed that the power drawn from other electrical devices is not correlated with the electrical device whose profile is derived. Further, it is assumed that the device profile is time-independent, i.e. that the steps in the program of the washing machine always take the same amount of time.

[0062] It is clear that not all of these assumptions hold all of the time in practice. But the principle of this embodiment has at least been demonstrated. More realistic techniques might be based on hidden Markov models as used in pattern recognition. In that case it is best to start from a known ideal device profile and identify states in the ideal device profile where the power draw or its first derivative is constant. Afterwards, the measurements can be used to identify probable actual states. It should also be noted that the co-location of the device steering 12 and the data collection 14, 16 units is of great value for the disaggregation algorithm as this conveys precise knowledge of the timing of the control commands to said disaggregation algorithm.

[0063] Fig. 7 shows a flow chart of an embodiment of a control method according to the present invention. The flow chart also shows which steps are carried out by an electrical device, by the control device and by the device profile optimizer. In step S10 the control device obtains and applies a device profile of an electrical device to control said electrical device. In step SI 2, the electrical device applies a program under control of the control device. In parallel the control device collects measurement data (step SI 4) and environmental data (step SI 6). The measured data are then (step SI 8) provided with tag information, whereafter (step S20) the measured data are sent to the device profile optimizer. The device profile optimizer analyzes the obtained data (step S22), retrieves a device profile of the electrical device to be optimized (step S24) and then (step S26) merges the retrieved device profile with the analyzed measurement data and analyzed environmental data.

[0064] Initial device profiles can be provided by a manufacturer that measures the response of an electrical device in a controlled laboratory setting. The electrical device being measured will be the only one connected to the laboratory test network and as such the device profile will not be influenced by other electrical devices. In real-life settings, however, this is not true anymore and e.g. line voltage is influenced by other electrical devices, both controlled and non-controlled, that are operating at the same time. The responses of all electrical devices are super imposed on each other. Furthermore, also external influences such as transformer tap changes have a distinct effect on a local electricity network.

[0065] Therefore it is preferred to process the measurements before they can be used to tailor the device profiles. Since the effects of external influences can be regarded as being random relative to the electrical device for which a refined profile is desired, statistical analysis is preferably used to filter out the random noise to obtain an actual device response.

[0066] A significant number of measurements is preferably obtained before an enhanced device profile is derived. Hence, in an embodiment measurements (of measurement data and/or of environmental data) of the same device models in different buildings/sites are pooled in order to both have access to more measurements more quickly and have measurements with less independent external influences that might otherwise bias the statistical analysis. In order to compare like with like however and to tailor profiles for specific circumstances, it is beneficial to enhance the measurements of the electrical responses with data describing the measurement and the environment in which it occurred, e.g. by use of measurement meta-data expressed as tags. For example, the date and time, outside temperature, the location of the measurement device (e.g. gauge) in the local electricity network, the device serial number etc. may be provided as additional meta-data along with the measured measurement data and/or environmental data.

[0067] Fig. 8 shows a schematic diagram of another embodiment of a control system 3 according to the present invention. In this embodiment it is shown that the control device 10, the device profile optimization device 60 and the profile repository 50 may be part of a cloud 4. In other embodiments only one or more of these elements may be part of a cloud. [0068] In general, the concepts of "virtual" and "cloud computing" include the utilization of a set of shared computing resources (e.g. servers) which are typically consolidated in one or more data center locations. For example, cloud computing systems may be implemented as a web service that enables a user to launch and manage computing resources (e.g. virtual server instances) in third party data centers. In a cloud environment, computer resources may be available in different sizes and configurations so that different resource types can be specified to meet specific needs of different users. For example, one user may desire to use small instance as a web server and another larger instance as a database server, or an even larger instance for processor intensive applications. Cloud computing offers this type of outsourced flexibility without having to manage the purchase and operation of additional hardware resources within an organization. A cloud-based computing resource is thought to execute or reside somewhere on the "cloud", which may be an internal corporate network or the public Internet. From the perspective of an application developer or information technology administrator, cloud computing enables the development and deployment of applications that exhibit scalability (e.g., increase or decrease resource utilization as needed), performance (e.g., execute efficiently and fast), and reliability (e.g., never, or at least rarely, fail), all without any regard for the nature or location of the underlying infrastructure.

[0069] In summary, the main advantages of the present invention are an improved precision of individual device profiles, accurate information from analysisfor updating device profiles, more predictable effects of demand-response device control, improved robustness of the electricity grid to unpredicted variations in energy production, and improved fitness of the electricity grid for renewable energy sources. If consuming devices (or batteries etc.) can be better scheduled better use can be made of renewable energy sources that are typically injected in the grid in a geographically distributed fashion. Thus, the electricity grid can accommodate the renewable energy sources better.

[0070] The invention has been illustrated and described in detail in the drawings and foregoing description, but such illustration and description are to be considered illustrative or exemplary and not restrictive. The invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims.

[0071] In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. A single element or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

[0072] A computer program may be stored / distributed on a suitable non- transitory medium, such as an optical storage medium or a solid-state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems.

[0073] Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. A device profile optimization device (60) for optimizing device profiles of one or more electrical devices (20, 22, 24), each device profile including energy usage-related and/or energy output-related information of an electrical device (20), comprising:

a measurement data input (61) that obtains measurement data of electrical parameters of an electrical device (20) to be optimized and/or of installations electrically coupled to said electrical device, and

an optimization unit (62) that optimizes the device profile by use of said measurement data into an optimized device profile.

2. The device as claimed in claim 1,

further comprising a processing unit (65) that processes said measurement data to derive additional energy usage-related and/or energy output-related information about the operation of the electrical device,

wherein said optimization unit (62) is configured to optimize the device profile by use of said measurement data and/or said additional energy usage-related and/or energy output- related data into the optimized device profile.

3. The device as claimed in claim 1,

further comprising a profile input (66) that obtains a device profile of an electrical device to be optimized, and a profile output (67) that provides the optimized device profile.

4. The device as claimed in claim 1,

wherein said optimized device profile is provided for use by said electrical device (20) and by electrical devices of the same type.

5. The device as claimed in claim 1 , wherein said measurement data include measurements of one or more parameters of an electrical voltage, electrical current and line voltage frequency of one or more of said electrical devices (20, 22, 24), a group of electrical devices, a building (40), a group of buildings including said electrical device, and an electrical network (42) to which said electrical device is connected.

6. The device as claimed in claim 1,

further comprising an environmental data input (6Γ) that obtains environmental data about the environment of said electrical device,

wherein said optimization unit (62) is configured to optimize the device profile by use of said measurement data and said environmental data into the optimized device profile.

7. The device as claimed in claim 1,

wherein said environmental data about the environment of said electrical device (20) include one or more of temperature, atmospheric pressure, humidity, date, time, location and weather conditions.

8. The device as claimed in claim 1,

wherein said optimization unit (62) is configured to perform a statistical analysis of said measurement data to obtain analyzed measurement data of said electrical device (20) for use in optimizing the obtained device profile.

9. The device as claimed in claim 8,

wherein said optimization unit (62) is configured to additionally use previously obtained measurement data to obtain said analyzed measurement data of said electrical device (20) by said statistical analysis.

10. The device as claimed in claim 8 or 9,

wherein said optimization unit (62) is configured to additionally use measurement data of electrical parameters of one or more electrical devices (20, 22, 24) of the same type as said electrical device (20) and/or installations (40, 42) electrically coupled to said one or more electrical devices (20, 22, 24) to obtain said analyzed measurement data of said electrical device (20) by said statistical analysis.

11. The device as claimed in claim 1 ,

further comprising a marking unit (63) that tags said measurement data and said optimized device profile with context data, in particular a relationship identifier indicating that the optimized device profile has been obtained by use of said measurement data.

12. The device as claimed in claim 10,

wherein said measurement data are tagged with a distance identifier indicating the distance, over an electrical network (42) to which said electrical device is connected, at which the respective measurement data are measurement from said electrical device (20).

13. The device as claimed in claim 1 ,

wherein said optimization unit (62) is configured to determine a measure of correctness of said optimized device profile.

14. A device profile optimization method for optimizing device profiles of one or more electrical devices (20, 22, 24), each device profile including energy usage-related and/or energy output-related information of an electrical device (20), comprising:

obtaining measurement data of electrical parameters of an electrical device (20) to be optimized and/or installations electrically coupled to said electrical device (20),

optimizing the device profile by use of said measurement data into an optimized device profile.

15. A control system comprising

a control device (10) that controls one or more electrical devices (20, 22, 24), one or more measurement data sensors (30, 32, 34, 36) that measure measurement data of electrical parameters of said one or more electrical devices (20, 22, 24) and/or installations (40, 42) electrically coupled to said one or more electrical devices (20, 22, 24),

a storage unit (50) that stores device profiles of said one or more electrical devices (20, 22, 24),

a device profile optimization device (60) as claimed in claim 1 for optimizing the device profile of one or more of said one or more electrical devices (20, 22, 24), each device profile including energy usage-related and/or energy output-related information of an electrical device (20),

wherein said control device (10) is configured to control an electrical device (20) by use of the corresponding optimized device profile.

16. The control system as claimed in claim 15,

wherein one or more of said measurement data sensors (30, 32, 34, 36) are configured to continuously monitor an electrical device (20) to be optimized and to measure measurement data of electrical parameters of said electrical device (20).

17. The control system as claimed in claim 15,

wherein one or more of said measurement data sensors (30, 32, 34, 36) are configured to measure measurement data of electrical parameters of an electrical device (20) to be optimized each time said electrical device (20) operates.

18. The control system as claimed in claim 15,

further comprising one or more environmental data sensors (70, 72) that measure environmental data about the environment of said one or more electrical devices (20, 22, 24).

19. The control system as claimed in claim 15,

further comprising a cloud (4) that includes at least one of the control device (10), the device profile optimization device (60) and said storage unit (50).

20. The control system as claimed in claim 15, further comprising one or more device sensors (80) that measure load characteristics of the electrical device.

21. A control method control system comprising

controlling one or more electrical devices (20, 22, 24),

measuring measurement data of electrical parameters of said one or more electrical devices and/or installations electrically coupled to said one or more electrical devices, storing device profiles of said one or more electrical devices (20, 22, 24), and optimizing the device profile of one or more of said one or more electrical devices as defined in claim 14, each device profile including energy usage-related and/or energy output-related information of an electrical device (20),

wherein an electrical device (20) is controlled by use of the corresponding optimized device profile.

22. A computer program comprising program code means for causing a computer to perform the steps of said method as claimed in claim 14 when said computer program is carried out on a computer.

23. A computer readable non-transitory medium having instructions stored thereon which, when carried out on a computer, cause the computer to perform the steps of the method as claimed in claim 14.