DETERMINATION OF A MODE OF OPERATION OF A BOILER

Abstract

This invention relates, but is not limited to, a method, a device, a computer program product and/or apparatus for determining the mode of operation of a boiler. The invention determines the mode of operation of a boiler through receiving a time series of electrical energy data of the boiler, determining energy events in the electrical energy data, determining a plurality of parameters of the energy events, and using an obtained model to determining the mode of operating of the boiler based on the obtained model and the plurality of parameters of the energy events. The invention further comprises a method of obtaining the model.

Claims

1-20. (canceled)

21. A method of building a model for detecting the mode of operation of boilers, the method comprising the steps of: receiving a plurality of training data sets for a plurality of boilers, each training data set comprising a time series of electrical energy data of the plurality of boilers; identifying an operation mode for each training data set; determining energy events in each training data set and computing a plurality of parameters of the energy events, the plurality of parameters comprising a number, a duration, and a regularity of the timing of the energy events; determining a measure of dependence between the identified operation mode and the determined plurality of parameters; and creating a model based on the determined measure of dependence, wherein the model is configured to assign a category indicative of the mode of operation of a boiler based on energy events in time series of electrical energy data from the boiler.

22. The method of claim 21, wherein the step of identifying an operation mode comprises at least one of: outputting to a human operator a visual representation of each of the training data sets; and receiving an input from the human operator indicative of the category; and obtaining control data from the plurality of boilers which identifies the operation mode.

23. The method of claim 21, wherein the model is a classification model, optionally wherein the classification model provides a confidence measure estimating the probability that the assigned category is accurate.

24. The method of claim 21, further comprising the step of determining a measure of confidence in the model, the measure of confidence dependent on a probability of accurately identifying the mode of operation of a boiler.

25. The method of claim 21, wherein the training data sets are obtained from a plurality of further boilers in a plurality of different locations, optionally where the training data sets are collected over multiple days of use of the plurality of further boilers.

26. The method of claim 21, wherein the determined modes of operation comprise at least one of an eco-mode and a pre-heat mode.

27. The method of claim 21, wherein the determined modes of operation comprise a mode in which energy usage takes place without providing domestic hot water or central heating demand.

28. The method of claim 21, wherein the regularity of the timing of the energy events comprises a standard deviation of the length of time between energy events.

29. The method of claim 21, wherein the plurality of parameters further comprises a first and/or last time of an energy event in the time series of electrical energy data.

30. The method of claim 21, wherein energy events are identified by an increase in the electrical energy data above a threshold or baseline of electrical energy usage.

31. The method of claim 21, wherein the time series of electrical energy data comprises at least one of current or voltage time series data.

32. The method of claim 21, wherein the step of determining energy events in the electrical energy data comprises determining energy events in a time window of the electrical energy data, optionally a daily time window.

33. The method of claim 21, wherein the mode of operation of the boiler is determined when a calculated probability of a correct determination reaches a threshold.

34. A method for determining a mode of operation of a boiler, the method comprising the steps of: receiving a time series of electrical energy data of the boiler, determining energy events in the electrical energy data, determining a plurality of parameters of the energy events, the plurality of parameters comprising a number, a duration, and a regularity of the timing of the energy events, obtaining a model for determining the mode of operation of a boiler, the model relating the mode of a boiler to a plurality of parameters determined from energy events in electrical energy data of boilers, and determining the mode of operating of the boiler based on the obtained model and the plurality of parameters of the energy events.

35. The method of claim 34, wherein the model is created by: receiving a plurality of training data sets for a plurality of boilers, each training data set comprising a time series of electrical energy data of the plurality of boilers; identifying an operation mode for each training data set; determining energy events in each training data set and computing a plurality of parameters of the energy events, the plurality of parameters comprising a number, a duration, and a regularity of the timing of the energy events; determining a measure of dependence between the identified operation mode and the determined plurality of parameters; and creating a model based on the determined measure of dependence, wherein the model is configured to assign a category indicative of the mode of operation of a boiler based on energy events in time series of electrical energy data from the boiler.

36. The method of claim 34, wherein the time series of electrical energy data is obtained from a current sensor associated with the boiler.

37. The method of claim 34, wherein the received model is a classification model and wherein the step of determining the mode of operation of the boiler comprises associating the boiler into one of a plurality of categories based on the classification model.

38. The method of claim 34, further comprising determining the mode of operation of the boiler based on a plurality of time series of electrical energy data of the boiler and determining the mode of operation of the boiler for each of the plurality of time series.

39. The method of claim 34, further comprising any one or more of the steps of: notifying a user of the mode of operation of the boiler, and/or sending a control signal to the boiler altering the state of the boiler, and/or prompting a user to alter the state of the boiler.

40. A system for determining a mode of operation of a boiler, the system comprising: a sensor for obtaining a time series of electrical energy data from the boiler; a processor configured to: identify energy events in the time series of electrical energy data from the boiler; determine a plurality of parameters of the energy events, the plurality of parameters comprising a number, a duration, and a regularity of the timing of the energy events; obtain a model for determining the mode of operation of a boiler, the model relating the mode of a boiler to a plurality of parameters determined from energy events in electrical energy data of boilers; and determine the mode of operation of the boiler based on the obtained model and the plurality of parameters of the energy events; and output means for outputting the mode of operation of the boiler to a user.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Preferred features of the present invention will now be described, purely by way of example, with reference to the accompanying drawings, in which:

[0055] FIG. 1 illustrates a system diagram of a monitoring system suitable for a boiler.

[0056] FIGS. 2a and 2b illustrates example sensor measurements for boilers in two different states.

[0057] FIGS. 3a, 3b and 3c illustrates example sensor measurements for boilers with complex states.

[0058] FIG. 4 illustrates a method of collecting classification data for a boiler.

[0059] FIG. 5 illustrates a method of constructing a classification system for a boiler mode of operation.

[0060] FIG. 6 illustrates interdependence between various parameters.

DETAILED DESCRIPTION

[0061] Boilers are commonly used to heat water in domestic environments. Modern boilers have a plurality of modes where each mode may affect the operating efficiency, cost, or eco-friendliness of a boiler's operation over a period of time. However, the mode of operation of a boiler is often hard to calculate or not visible to a user without investigating the internal operation of the boiler. This may be because the mode of operation of a boiler cannot be identified by a single operation, but is rather dependent on the boiler's operations over a period of time, such as how frequently it activates or the type of activations it performs. Therefore, it is advantageous to design a method and system to determine the mode of operation of a boiler based on externally available data.

[0062] In particular, the electrical energy supplied to the boiler can be used to distinguish between modes of operation such as pre-heat (where instant hot water is made available by regular heating events) and Time Proportional Integral (TPI, where the central heating system attempts to maintain a substantially constant home temperature). This is because the electrical energy signal can be analysed. For instance, the current signal can be analysed to determined operational events or energy events. Energy events can be identified in a number of ways. For instance peak detection algorithms could be used, energy usage above a threshold, average or base level could be considered an energy event, or another variation in the energy signal could indicate an event. Using parameters relating to the number, duration and regularity of the timing of these energy events the method can identify modes of operation, as they can distinguish between the regularity of modes such as pre-heat when compared to domestic hot water usage, or other modes which may have some regularity but differ in the expected number or duration of events.

[0063] Parameters relating to the number, duration and regularity of the timing of the energy events allow a strong classification or categorisation system to be built without making determination overly complication or computationally intensive. However, it is possible to add further parameters to the system such as the size or amplitude of the energy events. In further cases information relating to the flow in the hot water outlet, or the central heating system can also be used to add information to the system. The parameters may be used directly, but preferably a mathematical operation is used to understand the parameters. The mathematical operation may be a standard deviation, a maximum, a minimum, a count or an average (mean or median) or other means of obtaining a measure of the relative size, spread or variation of the energy events. In particular, the system may use any one or more of: [0064] a count of the number of energy events, [0065] a standard deviation of the duration of the energy events, [0066] a standard deviation of the interval between energy events and [0067] a measurement of the time when first activity was detected.

[0068] Further optional parameters include any one or more of: [0069] the timespan between first and last energy events, [0070] a measurement of the time when the last energy event occurred, [0071] the mean of the amplitude of the energy events, [0072] the mean duration of an energy event, [0073] the sum of the time of the energy events, [0074] the standard deviation of the energy used in the time period, [0075] the standard deviation of the amplitude of the energy events, [0076] the difference between the minimum and maximum electrical energy used, and [0077] the total electrical energy used across the period.

[0078] FIG. 1 shows a monitoring system for a boiler 104. The boiler 104 receives water from a domestic cold water supply 106 and produces a hot water output 105. The boiler is also connected to a central heating system through outlet and return pipes which may lead to a radiator (not shown). The skilled person would understand variations in the boiler design and operation are possible, including further features and boilers which do not feed radiators directly. The hot water output 105 may feed directly to a domestic supply or may feed into a hot water storage tank.

[0079] The monitoring system of FIG. 1 includes sensors attached to system components. Heating outlet sensor 111 and heating return sensor 112 may be temperature sensors configured to identify a change in temperature of the pipe indicating the presence of a flow. Similarly, hot water sensor 110 associated or attached to hot water output 105 may be a temperature sensor indicating flow of hot water from the boiler 104. The sensors 111,112,110 may alternatively be other sensors capable of measuring flow or related phenomena (e.g. temperature), able to indicate operation or amount of operation of the boiler or fluid flow. For instance, audio sensors, flow sensor or electromagnetic sensors may be used. Electrical energy sensor 114 is attached or associated with a power connection of the boiler 104 and is configured to detect the amount of electrical energy being used by the boiler. The electrical energy sensor 114 may be a current sensor or voltage sensor, or other suitable means for measuring electrical energy. Preferably the electrical energy sensor 114 is a current sensor or a voltage sensor. Preferably at least one, or all the sensors are attachable to the boiler, for example by clipping on to or grasping the wire/pipe they measure. In this way the system may easily be retrofitted to a boiler.

[0080] The sensors are preferably able to communicate with a controller 102. The controller 102 may be a microcontroller or other control device. It is preferably connected to the sensors by a connection means 113 which may be wired (e.g. a wire) or wireless (e.g. Bluetooth™ or.sup.Zigbee™) for each sensor. In many cases it is preferably wired to allow the controller to power the sensors. The controller 102 is preferably linked to a server 101 (either directly or through one or more intermediate devices), with which it may communicate wired or wirelessly. The server may also be in communication with a user device 103 this is typically using the internet, although it is also possible a user device would connect on a local area network (including to the controller or an intermediary device such as a hub). The user device 103, (for instance a personal electronic device such as a mobile telephone or smart, or other user interactive device) may allow a user to interact with the controller and/or to send instructions to the controller and/or to receive notifications from the controller, for instance regarding a diagnostic problem or a determined mode of operation. The notification may be by text message or notification in an application on the user device 103. The notification may prompt the user to take an action to control the boiler, or may inform the user of an action already taken by a boiler controller.

[0081] FIG. 2a shows example sensor readings or a boiler operating in an eco-mode mode of operation. In this mode of operation the boiler only operates when hot water demand exists (such as domestic hot water or heating). In the top part 201 of the figure the temperature (degrees Celcius) outputs are shown, where the domestic water 213 (measured by sensor 110) has peaks throughout the day. These will be caused user actions, such as showering, using the kitchen tap etc. Here 201 a single central heating peak is seen at about 3 pm in the flow and return signals 211, 212 (as measured by sensors 111, 112). The lower part of the figure shows a current plot (Amps). However, a voltage or other electrical signal plot may alternatively be used. The consumption spikes or energy events on the current plot (202) are correlated with the outlet temperature, with energy consumption resulting in hot-water usage or heating demand (where flow and return temperatures rise at the same time).

[0082] FIG. 2b shows a different mode of operation. In contrast to the eco-mode operation shown in FIG. 2a there are now a series of small current peaks throughout the day in the temperature plot 203. These do not appear to create a large increase in temperature of the overall system (e.g. limited effect on flow temperature and no visible effect on the domestic and return temperature), as none of the sensed temperatures rise above about 25 degrees, although current peaks are clearly seen in the plot 204. This is indicative of a pre-heat operative state. In pre-heat some amount of water is maintained at a high temperature so as to allow a quick response to any hot water request by a user (instead of a user having to wait a time period until the output water becomes hot enough for usage), this requires regular reheating due to temperature loss in the system. By using preheat the boiler is favouring comfort over cost reduction and low carbon emissions or efficiency. This operation state is often active by default on a boiler and the user may not be aware of it, for instance because it is an internal setting of the boiler. The pre-heat setting can be operated by heating water at a fixed time period, or activated when the water temperature drops below a threshold.

[0083] FIG. 3 demonstrates that the identification of the state of operation of the boiler is in many cases more complex than the relatively distinct examples shown in FIG. 2. The temperature plots 301 in FIG. 3a show a boiler which does not use pre-heat but has frequent consumption throughout the night, this results in many current peaks across the day of the current plot 302. Similarly, in FIG. 3b shows a boiler where small amounts of action are recorded in the temperature plot 303 overnight, with associated current spikes shown in the electrical plot 304. Plot 304 shoes a range of energy events, including energy events with different current peaks, or varying current peaks and sustained energy usage (see hour 18). Finally, a central heating system may be configured to provide heating throughout the day. FIG. 3c demonstrates a series of temperature spikes throughout the time period of the temperature plot 305, with corresponding spikes shown in the current plot 306. This variation in the operation of different systems, and user impacts through demand for domestic water make it difficult to easily identify the operation state of the boiler. This is increased further when a mix of modes of operation is used, such as pre-heat and central heating demands.

[0084] FIG. 4 shows an example collection of boiler data. In a first step time-series data 401 is collected or otherwise obtained. This may be obtained from a plurality of systems as described in FIG. 1, or otherwise measured. FIG. 4 shows collection of both water temperature data and electrical signal data (in this case current), however in a particular case only the current (A) data is used. The current data of the whole period may be used, but preferably a selected or pre-set portion of the time data is used. This may be selected by the controller or a programmer of the system. Preferably the time window is a period of typically low use of domestic hot water, such as between midnight and 5.30 am, however other time windows are possible. The time window should be of a reasonable length so as to capture a range of operation, preferably at least 4 or 5 hours. Once the time series data has been captured the variables or parameters can be extracted 402 from the time series. The parameters can then be put into a classification system or method 403 to determine the operation state of the boiler.

[0085] FIG. 5 shows the classification system may be built using machine learning or classification analysis. For instance, using a random forest algorithm, although other algorithms will be known such as neural nets or support vector machines. A random forest algorithm uses a training set of data to build a set of decision trees or classification trees (typically sampling from the training set for each decision tree and selecting a random subset of features at each step of the tree to create a range of decision trees which will average to an accurate classification). There may be any number of decision trees used in the random forest, for instance 100 or 500. When classifying a new boiler, the data is fed into the set of trained decision trees and the operating mode is determined based on a summary, average or vote from each of the outcomes of each of the trained decision trees. The classification system can take as inputs the selection of parameters and use a training process to identify a suitable classification based on the selected parameters. The system may break down the space of the parameters into subregions, with each subregion representing a mode of operation of the boiler.

[0086] The classification system uses training data, this data is collected from a range of other boilers 501 and may be collected across multiple different days and/or boilers. Each data set used in the training data related comprises a time series of current data for a time period, in this case time period between midnight and 5.30 am. The training data may include multiple days of the same boiler in the same of different modes, or multiple different boilers. This training data may optionally be sampled 502 to reduce memory or computation requirements or to avoid replication of highly similar data. The training data is identified, by annotating 503 the mode of operation, for instance through human intervention, by tags, annotations or otherwise further data included with the time series of current data, or more complex identification means. The time series of current data is then analysed or processed to identify energy events, relating to peaks or fluctuations from a base level or threshold in the time series data. This process is similar to a single boiler example. The relevant parameters, variables or numerical representations of the relationships of and between energy events is then be extracted 504. The training data can then be used to train 505 a classifier based on the parameters, such as by the random forest method discussed above, or other classification method. This model is then able assign a category indicative of the mode of operation of a further or test boiler based on time series current data from that further or test boiler. The training of the classification system may be relatively computationally expensive, but can advantageously allow for a very simple classification of new data.

[0087] When classifying a boiler the controller 102 may receive current data from sensor 114. It may collect the current data over the required time period, or plurality of time periods. The controller 102 may then send this data to the server 101 or process it locally. The processing requires receiving or obtaining the classification model, which may be stored on the controller, or on the server or downloadable from the server. The controller (or server) can then classify the current data for the or each time period based on the classification model, obtaining a determined state of operation for the or each time period. In some embodiments the processing may be shared between the controller and/or server or other computation device, for instance the controller may identify energy events and the server may perform the classification.

[0088] Where multiple time periods are used the controller can perform a check that the same, or expected mode of operation has been calculated in each period, or if a confidence threshold or probability is not achieved (for instance because the mode of operation has been determined to be different 20%, 50% or 70% of the time periods), delay a determination until further current data is available. Where single or multiple time periods are used the controller can calculate a confidence or probability in the accuracy of the determination, this may be calculated by the confidence of the model in the answer (e.g. how many, or the percentage of the random forest decision trees predict the determined mode of operation). It is possible that the confidence in the accuracy of each prediction and the number of predictions based on different time periods are mathematically combined to form an overall confidence score or accuracy probability for multiple time period measurements. Once a determination of the mode of operation of the boiler has been made this can be notified to the user through user device(s) 102. For instance, this may be in the form of a notification or a prompt to change the mode of operation. Alternatively, the controller 102 or server 101 may automatically send an instruction to change the mode of operation of the boiler if this is available (based on the available control of the boiler and previous user instructions).

[0089] The server will typically include other conventional hardware and software components as known to those skilled in the art. While a specific architecture is shown by way of example and specific software technologies and vendors have been mentioned, any appropriate hardware/software architecture may be employed. Functional components indicated as separate may be combined and vice versa. For example, the functions of server 101 may in practice be implemented by multiple separate server devices, e.g. by a cluster of servers.

[0090] FIG. 6 shows graphs of correlations between the parameters or numerical features of the events of the boiler time series data. The data sets used to create the plots have been created from data recorded about a plurality of boilers in houses, typically over a plurality of time periods, where approximately half of the boilers use a pre-heat mode (labelled True) and the remaining half use an eco-mode (labelled False). Each point on each graph represents a data set from a training period (in this case between midnight ant 5.30 pm). Parameters relating to the duration of energy events (in this case the standard deviation), the time between energy events (standard deviation), the number of energy events and the time of the first event occurring (in hours) are shown. A log function has been used to allow simpler visualisation of the parameters. Each of these parameters may have been available from the current time series of the boiler.

[0091] The first row of graphs 601, 602, 603, and 604 show the standard deviation of the duration of energy events on the y-axis. The second row of graphs 605, 606, 607, and 608 show the standard deviation of the time between energy events. The third row of graphs 609, 610, 611 and 6112 show the number of energy events and the final row of graphs 613, 614, 615 and 616 shows the time of the first energy event (in hours from the start of the time period). In each graph dense outliers may represent default values where no information is available (e.g. values of 10 in first h). The horizontal x axes show the standard deviation of the duration of energy, the standard deviation of the time between energy events, the number of intervals and the time of the first energy event respectively. This means that the matrix of plots shows the relationships between each of the parameters in turn. The diagonal plots 601, 606, 611 and 616 are shown as histograms (as they would otherwise plot the same data on both axis) which demonstrate the relationship between the two modes for that variable. It is visible from the plots that although some differentiation is possible for each of the parameters, none is able to clearly distinguish the modes. For instance, the pre-heat mode shows less variation in the length of intervals in plot 606 and a more consistent number of intervals in plot 611.

[0092] It will be understood that the present invention has been described above purely by way of example, and modification of detail can be made within the scope of the invention.

[0093] The above embodiments and examples are to be understood as illustrative examples. Further embodiments, aspects or examples are envisaged. It is to be understood that any feature described in relation to any one embodiment, aspect or example may be used alone, or in combination with other features described, and may also be used in combination with one or more features of any other of the embodiments, aspects or examples, or any combination of any other of the embodiments, aspects or examples. Furthermore, equivalents and modifications not described above may also be employed without departing from the scope of the invention, which is defined in the accompanying claims.