GB2571145A

GB2571145A - Flow alert system and method

Info

Publication number: GB2571145A
Application number: GB1807254.6A
Authority: GB
Inventors: Parson Oliver; Maciol Ryszard
Original assignee: Centrica PLC
Current assignee: Centrica PLC
Priority date: 2018-05-02
Filing date: 2018-05-02
Publication date: 2019-08-21
Anticipated expiration: 2038-05-02
Also published as: GB2571145B; GB201807254D0; WO2019211618A1

Abstract

A method for filtering alerts in response to flow in a fluid conduit is described. Alerts are triggered in response to detection of flow, and the method comprises the steps of receiving a set of historical flow data relating to flow in the fluid conduit for a first time period. Identifying one or more recurring flow patterns based on the set of historical flow data. Designating one or more alert suppression periods based on the identified recurring patterns. Receiving subsequent flow data relating to flow in the fluid conduit for a second time period. Detecting flow events in the subsequent flow data to generate one or more protentional flow alerts and suppressing potential flow alerts for flow events during the designated alert suppression periods within the second time limit. Later embodiments relate to computer readable medium comprising software to perform said method and apparatus and a system for carrying out said method.

Description

FLOW ALERT SYSTEM AND METHOD

The present application relates to systems and methods for alerting a user about fluid flows. In particular, the application relates to filtering flow alerts about fluid flow in a conduit, for example flow in a water or gas pipe.

There exist systems and methods for monitoring fluid flow in a conduit, which may alert a user about anomalous flow (e.g. flow that may appear unintentional and/or be indicative of a leak). Such systems may identify (continuous) flow detected for more than a threshold time period as anomalous, and thus generate a user alert. False alarms may occur where intentional fluid flow is longer than a threshold time limit. Users who regularly use the fluid for long periods (e.g. frequent long showers) may receive false alarms frequently, which can be frustrating.

The inventors have appreciated that there would be a benefit in identifying patterns in fluid use which may identify time periods during which intentional flows that may trigger alerts are likely to occur. By suppressing alerts during such time periods it may be possible to reduce the number of false alarms.

Aspects of the invention are set out in the independent claims and preferable features are set out in the dependent claims.

There is described herein a method for filtering alerts in response to flow in a fluid conduit, wherein alerts are triggered in response to detection of flow, the method comprising the steps of: receiving a set of historical flow data relating to flow in the fluid conduit for a first time period; identifying one or more recurring flow patterns based on the set of historical flow data; designating one or more alert suppression periods based on the identified recurring flow patterns; receiving subsequent flow data relating to flow in the fluid conduit for a second time period; detecting flow events in the subsequent flow data to generate one or more potential flow alerts; and suppressing potential flow alerts for flow events during the designated alert suppression periods within the second time period.

By suppressing alerts during alert suppression periods it is possible to reduce the likelihood of false flow alarms occurring. Advantageously, because the alert suppression periods are identified based on historical flow data for the fluid conduit concerned, alert suppression can reduce false alarms more effectively, whilst still allowing users to be alerted about flow events occurring at times where (extended) intentional flow is less likely.

Generally the alert suppression periods are recurring time intervals, for example periods that recur at the same time every day or every week. In some cases the length of time between consecutive recurring time intervals is not constant, e.g. the timing of the alert suppression periods varies dependent on the day of the week, for example a first set of alert suppression periods recurring every weekday (e.g. Monday to Friday) may differ from a second set of alert suppression periods recurring every weekend day (e.g. Saturday and Sunday).

Suppressing potential flow alerts may involve detecting flow events in the subsequent flow data and then suppressing alerts for flow events detected during suppression periods. Alternatively, suppressing potential flow alerts may involve not detecting flow events during suppression periods (e.g. not analysing the flow data for suppression periods), and thus the potential flow alerts are not generated for the suppression periods.

Preferably, identifying a recurring flow pattern comprises: dividing the first time period into time intervals; and grouping the time intervals into a plurality of groups of temporally recurring time intervals; and designating one or more alert suppression periods is based on: a first measure for each group that is independent of flow data for time intervals in the remainder of the plurality of groups; and/or a second measure for each group that is dependent on flow data for time intervals in at least two of the plurality groups.

The first and/or the second measure may be indicative of the number of flows identified in intervals in each group and/or may be representative of the probability (or likelihood) of a flow occurring in each group (e.g. based on historical data).

Preferably the time intervals are regular, e.g. of identical length. In some examples the time intervals are each an hour. The temporally recurring time intervals in each group may recur, for example, at the same time each day, or the same time each week. Generally the time intervals in each group recur at the same rate, for example periods that recur at the same time every day or every week are grouped together. In some cases the length of time between consecutive recurring time intervals in a group is not constant, e.g. time intervals are grouped dependent on the day of the week, for example a first set of time intervals recurring at the same time every weekday (e.g. Monday to Friday) may be grouped into a first group, different from a second group for a second set of time intervals recurring at that same time every weekend day (e.g. Saturday and Sunday)

The first measure for each group may be based on flow data for time intervals in that group (and does not take into account flow data for time intervals not in that group, or in other groups). The first measure is an absolute measure, e.g. a measure that is agnostic of flow at time intervals in other groups. Preferably the second measure for each group is based on flow data for time intervals across all the historical data, e.g. for time intervals in all groups of the plurality of groups. The second measure is a relative measure, e.g. a measure of how flows for time periods in each group compare to flows in other groups.

Preferably the method further comprises, after dividing the first time period into time intervals, obtaining (e.g. calculating, or otherwise determining or receiving) flow characteristic data for each time interval from the historical flow data. This may involve identifying “flow events”, e.g. periods of substantially continuous flow, from the flow data. Flow events may be interspersed by “no flow” periods, where no flow is detected in the fluid conduit. In some cases a flow event is a continuous flow at a substantially constant rate. In some embodiments the flow may be considered to be at a substantially constant rate if the flow rate does not change by more than a certain rate change proportion of its starting value. The rate change proportion may be at least 2%, preferably at least 5%, more preferably at least 10%, or around 20%. The rate change proportion may additionally or alternatively be less than 60%, preferably less than 40%, more preferably less than 30%.

The flow characteristic data may comprise one or more of: an indication of whether or not there was flow (e.g. whether there was any flow at all over the time interval, or whether at least one “flow event” is identified in the interval) in the fluid conduit in the time interval; a duration of flow in the fluid conduit during the time interval (e.g. the time, in minutes or seconds, in the time interval for which there was some flow in the conduit); a rate or average (e.g. mean) rate of flow in the fluid conduit during the time interval; a number of flow events detected in the time interval, e.g. the number of distinct flow events (and potentially the duration of each flow event); and a flow volume of flow passing through the conduit in the time interval. An indication of whether or not there was flow could be positive if there was at least one flow of more than a predetermined length in the interval, in some embodiments this may be an indication of whether there was a substantial flow in the interval, e.g. a flow of more than a substantial duration threshold. The number of flow events detected may be flow events of more than a predetermined length. The predetermined length/duration may be, e.g. at least 3 or 5 seconds and/or not more than 30 minutes, not more than 20 minutes or not more than 10 minutes. In some embodiments the predetermined length may be at least 5 seconds and not more than 30 seconds. The substantial duration threshold may be at least 2 minutes, at least 5 minutes, at least 10 minutes, or around 15 minutes. The substantial duration threshold may be less than 40 minutes, less than 30 minutes or preferably less than 15 minutes. For example the substantial duration threshold may be between 10 and 20 minutes.

The step of identifying one or more recurring flow patterns may comprise detecting flow events from the historical flow data.

Preferably, designating the one or more alert suppression periods comprises designating recurring time intervals corresponding to the groups of temporally recurring time intervals which satisfy both: a first criterion based on the first measure; and a second criterion based on the second measure. By looking at an absolute measure (e.g. the first measure) in conjunction with a relative measure (e.g. the second measure), the alert suppression periods may be more effective. For example, by making comparisons between groups it may be possible to avoid classing too many periods as suppression periods e.g. if there is already a constant leak. Equally, by considering the time periods in one group independently it may be possible to avoid classing too many periods as suppression periods in the case of very little data (e.g. if a large proportion of flows happen to occur in time periods in the same group, but there are not enough flows for this to be statistically significant).

Preferably the first time period spans multiple days, preferably at least one to two months and less than 12 months. In some embodiments the first time period may be at least three weeks and not more than 6 months. The duration of the first time period should be selected to be long enough to provide sufficient data to determine the difference between genuine patterns and noise. It is also desirable that it is short enough to take into account recent changes to patterns in flow, e.g. caused by recent changes in a customer/user’s routine.

Preferably, the first measure is based on (or corresponds to) the proportion of time intervals in the group for which flow (e.g. at least one flow event) is detected, or the probability of flow occurring in the time interval, based on the historical flow data. The first measure may be indicative of the proportion of intervals in the group having at least one flow (e.g. a distinct flow event) or some flow (e.g. a part of a distinct flow event occurs during the time interval). Optionally the first measure may be indicative of the proportion of intervals in the group having one or some flow of more than a predetermined length or a substantial flow of more than a substantial duration threshold. The predetermined length/duration may be, e.g. at least 3 or 5 seconds and/or not more than 30 minutes, not more than 20 minutes or not more than 10 minutes. In some embodiments the predetermined length may be at least 5 seconds and not more than 30 seconds. The substantial duration threshold may be at least 2 minutes, at least 5 minutes, at least 10 minutes, or around 15 minutes. The substantial duration threshold may be less than 40 minutes, less than 30 minutes or preferably less than 15 minutes. For example the substantial duration threshold may be between 10 and 20 minutes.

The first measure may be based on the proportion (or number) of intervals in the group having at least a threshold number of flows, or distinct flow events, (optionally only flows of more than the predetermined length or substantial duration threshold, as above). The threshold number of flows may be at least two, three or four flows. In some embodiments the threshold number of flows may be at least five or at least 10.

Preferably, the first criterion is satisfied if the proportion of time intervals in the group for which flow is detected is greater than a first threshold proportion. Thus it may be possible to avoid classifying a time interval as a suppression period if there are very few flows in the historical data, but those flows all happen (e.g. by random chance) to be in the same recurring time interval (e.g. at the same time of day).

For example, the first threshold proportion may be at least 1%, preferably at least 2% or 3%, more preferably around 5%. Additionally or alternatively the first threshold proportion is less than 30%, preferably less than 20%, more preferably less than 10%. In some embodiments the first threshold proportion is at least 3% and not more than 8%. The first threshold proportion may be set high enough so as not to class too many time intervals as suppression periods, but low enough so that suppression periods are still detected where appropriate.

The first threshold proportion may be selected in dependence on the length of the time periods, e.g. because shorter time periods are less likely to have any flows occur. A first threshold of between 3% and 10% has been found to work well when time periods are between 30 minutes and 2 hours in length, e.g. a first threshold proportion of around 5% when the time intervals are around 1 hour.

Preferably the second measure is based on, or corresponds to, the flows detected in the time intervals in the group in question as a proportion of the total flows detected in time intervals in at least two groups (the at least two groups including the group in question and one other group). In some embodiments the second criterion is satisfied if the proportion of total flows (or distinct flow events) in the at least two groups (or in all the groups, e.g. in the entire second period) that are detected in the group in question is greater than a second threshold proportion. Thus it may be possible to prevent all (or a large proportion of) time intervals being classified as suppression periods in the case of a constant (or almost constant) leak.

In some embodiments the second measure is indicative of the intervals in the group in question having at least one flow (e.g. a distinct flow event) or some flow (e.g. a part of a distinct flow event occurs during the time interval) as a proportion of the intervals in the at least two groups which have at least one flow or some flow. Optionally the second measure may be indicative of the proportion of intervals in the group and the at least two groups having one or some flow of more than a predetermined length or a substantial flow of more than a substantial duration threshold. As before, the predetermined length/duration may be, e.g. at least 3 or 5 seconds and/or not more than 30 minutes, not more than 20 minutes or not more than 10 minutes. In some embodiments the predetermined length may be at least 5 seconds and not more than 30 seconds. The substantial duration threshold may be at least 2 minutes, at least 5 minutes, at least 10 minutes, or around 15 minutes. The substantial duration threshold may be less than 40 minutes, less than 30 minutes or preferably less than 15 minutes. For example the substantial duration threshold may be between 10 and 20 minutes.

The second measure may be based on the intervals in the group having at least a threshold number of flows, or distinct flow events (optionally only flows of more than the predetermined length or substantial duration threshold, as above) as a proportion of the intervals in the at least two groups having at least the threshold number of flows. The threshold number of flows may be at least two, three or four flows. In some embodiments the threshold number of flows may be at least five or at least 10.

The second criterion may be satisfied if the proportion of intervals having at least one (or some) flow (or more than a threshold number of flows) in the at least two groups (or in all the groups, e.g. in the entire second period) that are detected in the group in question is greater than a second threshold proportion.

Preferably, the second threshold proportion is at least 3%, preferably at least 4%, more preferably at least 8%, preferably around 10%. Preferably, the second threshold proportion is less than 40%, preferably less than 30%, more preferably less than 20% or 15%. In some embodiments the second threshold proportion is at least 8% and not more than 15%.

The second threshold proportion may also be selected in dependence on the length of the time periods, e.g. because shorter time periods are less likely to have a large proportion of the total flow. A second threshold proportion of at least 8% and not more than 15% has been found to work well for time periods between 30 minutes and 2 hours in length, e.g. a second threshold proportion of around 10% when the time intervals are around 1 hour.

Generally the plurality of groups comprises at least two groups, e.g. for at least two recurring time periods. Preferably the plurality of groups comprises at least three groups, or more preferably at least 10 or at least 20 groups. Generally less than 50 groups is desirable. E.g. there may be one group corresponding to each hour of the day, i.e. 24 groups.

In some embodiments the plurality of groups comprises at least two groups that are determined independently for different days of the week, for example being determined on weekdays independently of weekends.

In certain embodiments, designating the one or more alert suppression periods comprises designating recurring time intervals that meet one or both of the following criteria: flow is identified in the interval for at least a first threshold proportion of days in the set of historical data; and flows identified in the interval exceed a second threshold proportion of total flows identified in the set of historical data.

Suppressing potential flow alerts may comprise changing one or more criteria for triggering an alert, such as increasing an alert threshold. For example, where an alert is triggered by detecting a flow of longer than a threshold flow duration, increasing the alert threshold may comprise increasing the threshold flow duration required to trigger an alert, e.g. applying a longer flow duration threshold during the alert suppression periods.

Preferably, suppressing potential flow alerts comprises suppressing flow alerts entirely. For example, this may be by preventing alerts being generated in the first place, e.g. the analysis to determine whether the one or more criteria for triggering an alert is met may not be performed during suppression periods, or the flow data or measurements used in such analysis may not be made. This could have power saving advantages. Alternatively, flow events may continue to be detected and potential alerts generated, but during the suppression periods such potential alerts are not converted to actual alerts (e.g. alert messages), or alert messages may not be sent to or displayed to a user.

In some embodiments the method further comprises: updating the designation of the alert suppression periods using the subsequent flow data. For example, the step of identifying one or more recurring flow patterns may be repeated based on the set of subsequent flow data (in conjunction with some/all of the historical flow data), and the step of designating one or more alert suppression periods may be performed based on these newly identified recurring flow patterns. It may be preferable to use all the historical data, e.g. since using a larger data set may give more accurate results, or it may be preferable to use only the most recent data (e.g. data for flows in the previous two to three weeks or two to three months) in order to make adjustments as user behaviour changes. If using the most recent data, this may include subsequent flow data and a portion of the historical flow data.

Preferably updating the designation of the alert suppression periods is performed periodically, e.g. at periodic time intervals such as every day, or every week.

Preferably, the alert suppression periods and/or the time intervals are each at least 10 minutes, preferably at least 20 minutes, more preferably at least 30 minutes, or preferably around one hour. Additionally or alternatively the alert suppression periods and/or the time intervals are preferably each less than around 4 hours, preferably less than three hours or two hours, more preferably around one hour. In some embodiments, the alert suppression periods and/or the time intervals are between 40 minutes and 90 minutes.

Preferably the potential flow alerts are generated in response to the detection of fluid flowing in the fluid conduit substantially continuously for a minimum flow time. Fluid flowing in the conduit can be detected from the subsequent flow data. In some embodiments, potential flow alerts are generated in response to the detection of fluid flowing in the fluid conduit for the minimum flow time at a rate between a lower flow rate threshold and an upper flow rate threshold. Thus flow alerts will be generated when flow is at a substantially constant rate (i.e. between the two rate thresholds). Flow for a long period of time at the same rate may be more likely to indicate unintentional flow (e.g. a leak) than if the flow rate varies.

In some embodiments the minimum flow time period is at least 2 minutes, preferably at least 5 minutes or 10 minutes, more preferably around 15 minutes. Alternatively or additionally the minimum flow time period is less than an hour, preferably less than 30 minutes or 20 minutes. Preferably the minimum flow time period is between around 10 and 20 minutes.

In some embodiments, the minimum flow time required to generate a potential flow alert is selected by a user, or is based on historical flow data. This may further prevent false flow alerts, e.g. if a user generally uses the fluid (e.g. water) for a long period of time.

In some embodiments, the one or more potential flow alerts are each generated in response to detection of a flow event that is characteristic of unintentional flow. Unintentional flow may be indicative of a leak.

In some examples, suppressing potential flow alerts comprises: for each potential flow alert, establishing whether the potential flow alert has been detected during one of the one or more designated alert suppression periods; and upon establishing that the potential flow alert has been detected during one of the one or more designated alert suppression periods, preventing notification of a flow alert to a user.

In other examples, suppressing potential flow alerts comprises: discarding or ignoring at least some or all of the received subsequent flow data relating to the one or more designated alert suppression periods prior to detecting flow events in the subsequent flow data to generate one or more potential flow alerts. Thus essentially the step of suppressing potential flow alerts may be performed prior to the step of detecting flow events in the subsequent flow data. This may reduce processing requirements of the system.

Preferably, the set of historical flow data and/or the subsequent flow data comprises one or more of: an indication of whether there is a flow in the fluid conduit; a duration of a flow in the fluid conduit; the temperature of the fluid conduit and/or the ambient temperature surrounding the fluid conduit; and a rate of flow in the fluid conduit. In other embodiments, it may be possible to determine one or more of: an indication of whether there is a flow in the fluid conduit; a duration of a flow in the fluid conduit; the temperature of the fluid conduit and/or the ambient temperature surrounding the fluid conduit; and a rate of flow in the fluid conduit from the historical flow data and/or the subsequent flow data.

Preferably, the first time period is at least one week, preferably at least two weeks, more preferably around 3 weeks and/or preferably less than 2 months, or less than one month.

Preferably, the method further comprises alerting a user based on a potential flow alert generated outside the designated alert suppression periods. The step of alerting a user may comprise outputting a flow alert notification to a user interface; and/or sending a flow alert notification to a user device. The user device may be associated with an operator or other user of the fluid conduit. Alerting the user may be achieved by displaying the alert on a user interface, e.g. a user interface of a flow monitor device, generating a sound, e.g. by the flow monitor device, sending a message, such as an email, SMS message or notification for an application to a user device (e.g. a mobile phone, tablet or computer).

Preferably, receiving a set of historical flow data relating to flow in the fluid conduit for a first time period and/or receiving subsequent flow data relating to flow in the fluid conduit for a second time period comprises making a measurement indicative of flow in the fluid conduit. For example, the measurement may be of a quantity or characteristic relating to flow of fluid in the conduit. The measurement may be a direct measurement of flow (e.g. from a flow meter, such as a mechanical, magnetic or ultrasonic flow meter) or may be e.g. a temperature which is indicative of flow in the conduit.

Preferably, measuring flow in the fluid conduit comprises: measuring a first temperature being the temperature of the outer surface of the fluid conduit; measuring a second temperature being the ambient temperature outside of the fluid conduit; and making a flow determination by determining a measure of the convergence of the first and second temperatures over time.

There is also described herein a computer-readable medium comprising software code adapted, when executed on a data processing apparatus, to perform a method substantially as set out above.

There is also described herein an apparatus for filtering alerts in response to flow in a fluid conduit, wherein alerts are triggered in response to detection of flow, the apparatus comprising: a memory; an interface for receiving flow data relating to flow in the fluid conduit; and a processor operable to: receive a set of historical flow data relating to flow in the fluid conduit for a first time period; identify one or more recurring flow patterns based on the set of historical flow data; designate one or more alert suppression periods based on the identified recurring flow patterns; receive subsequent flow data relating to flow in the fluid conduit for a second time period; detect flow events in the subsequent flow data to generate one or more potential flow alerts; and suppress potential flow alerts for flow events during the designated alert suppression periods within the second time period. This apparatus may be a flow monitoring device to be provided at or near the fluid conduit, or it may be a server at a remote site which is capable of receiving flow data from a flow monitor at or near the fluid conduit. The flow monitoring device may be a flow meter, such as a mechanical, magnetic or ultrasonic flow meter, or may monitor flow by looking at temperature differences.

There is also described a system for filtering alerts in response to flow in a fluid conduit, comprising: an apparatus for filtering alerts in response to flow in a fluid conduit and one or more sensors operable to make a measurement relating to flow in the fluid conduit (e.g. a measurement of a quantity/characteristic indicative of flow in the conduit). Generally the one or more sensors are configured for placement at or near the fluid conduit (e.g. for coupling directly to the fluid conduit or for positioning near the fluid conduit for making measurements in the vicinity of the conduit). In some embodiments the one or more sensors are integrated into a flow monitor device.

Preferably, the one or more sensors comprise one or more temperature sensors, preferably the one or more sensors comprise: a first temperature sensor arranged to be mounted on the outer surface of the fluid conduit to measure a first temperature being the temperature of the outer surface and arranged to generate a first temperature signal indicative of the first temperature; and a second temperature sensor arranged to be positioned spaced apart from the first temperature sensor, to measure a second temperature being the ambient temperature outside of the fluid conduit and to generate a second temperature signal indicative of the second temperature.

Preferably, the system further comprises a fluid conduit.

Preferably there is also provided a user interface configured to present a flow alert to a user, such as a screen, for example a screen of a smartphone, personal computer, smart watch or the like, or a speaker, e.g a loudspeaker or buzzer for sounding an alert.

Any system feature as described herein may also be provided as a method feature, and vice versa. As used herein, means plus function features may be expressed alternatively in terms of their corresponding structure.

Any feature in one aspect of the invention may be applied to other aspects of the invention, in any appropriate combination. In particular, method aspects may be applied to system aspects, and vice versa. Furthermore, any, some and/or all features in one aspect can be applied to any, some and/or all features in any other aspect, in any appropriate combination.

It should also be appreciated that particular combinations of the various features described and defined in any aspects of the invention can be implemented and/or supplied and/or used independently.

BRIEF DESCRIPTION OF THE FIGURES

Methods and systems for flow alerts are described by way of example only, in relation to the Figures, wherein:

Figure 1 shows an exemplary system for implementing flow alerts;

Figure 2 shows an exemplary method for implementing flow alerts;

Figure 3 shows an exemplary method for identifying recurring flow patterns and designating alert suppression periods;

Figure 4 is a graph showing exemplary flow data illustrating the probability of a flow occurring independently by time of day;

Figure 5 is a graph showing exemplary flow data illustrating the probability of a flow occurring, given that a flow will occur at some point; and

Figure 6 shows an exemplary arrangement of a flow monitor device.

DETAILED DESCRIPTION

Figure 1 shows an exemplary system 100 for detecting flow conditions and providing flow alerts. The system 100 comprises a flow monitor 110 located at a user’s premises. The premises may be a home, or residential premises, or could be a commercial premises, such as an office, shop or factory.

The flow monitor 110 is in communication with a first temperature sensor 112 and a second temperature sensor 114. This may be via wired or wireless communication.

The flow monitor 110 is connected wirelessly to the user’s local network I internet access infrastructure, for example, to a wireless or wired home router I access point 150, which in turn provides access to the Internet 170 through a modem 160, such as an ADSL or fibre modem. Depending on access technology, router 150 and modem 160 may be combined in a single device or replaced with other access devices appropriate to the access technology.

The first temperature sensor 112 is arranged for locating on the surface of a fluid conduit 120. The first temperature sensor 112 thus provides a temperature reading indicative of the temperature of the fluid in the conduit 120. The fluid conduit 120 may be, for example, a water pipe, or alternatively a gas or oil pipe.

The second temperature sensor 114 is arranged to sense an ambient temperature, e.g. the temperature of surrounding air in the vicinity of the fluid conduit 120. The first temperature sensor 112 and the second temperature sensor 114 are arranged to send temperature readings to the flow monitor 110. In some embodiments the flow monitor 110 may also be configured to send instructions to the first temperature sensor 112 and the second temperature sensor 114. For example, the instructions may concern the rate of taking temperature measurements, the rate of communication of those temperature measurements and/or may switch on/off (or activate/deactivate) the sensors 112, 114. Thus, it may be possible to conserve energy and battery power.

The fluid temperature, e.g. measured by first temperature sensor 112, is generally different from the ambient temperature when fluid is flowing in the conduit 120. When no fluid is flowing in the conduit 120 the conduit and ambient temperatures tend to converge. Thus flow of fluid in the conduit 120 can be detected by non-convergence of the temperatures measured by the first temperature sensor 112 and the second temperature sensor 114. Examples of methods and apparatus for detecting flow based on the convergence of temperatures can be found in UK patent no. GB2536364, International publication no. W02018/007802 and UK patent application no. GB1718195.9.

The flow monitor 110 receives temperature readings from the first 112 and second 114 temperature sensors and uses this to identify when fluid is flowing in the conduit. Temperature may be measured by the sensors 112, 114 at a certain sampling rate, e.g. every 3 seconds, every 5 seconds, every 10 seconds, every 20 seconds or every minute.

The flow monitor 110 may record and store information relating to flow in the conduit as flow data. Flow data may comprise an indication of flow in the conduit at each sampling time throughout the first time period. The sampling times could, for example, mirror the sampling rate of the temperature measurements, e.g. every 3 seconds, every 5 seconds, every 10 seconds, every 20 seconds or every minute. The indication of flow can be an indication of whether or not there was flow, e.g. a binary flag (flow/no flow). Flow data can optionally include the rate of fluid flow. The flow monitor 110 may classify a substantially continuous period of flow as a flow event, and may store information such as the start and end times, duration and rate of flow for each flow event. The rate of flow for a flow event may be an average rate of flow for the period of the flow event. In some embodiments, the flow data includes the raw temperature measurements from the first and second temperature sensors.

The flow monitor 110 can use the flow data to identify flows that are likely to be unintentional or indicative of a leak. From this the flow monitor 110 is capable of generating flow alerts based on the flow data. The flow monitor 110 may also use the flow data to identify patterns in usage, which may be useful for suppressing flow alerts. For example, the flow monitor may determine times of day (or week) when flow is most likely to occur. Further explanation about identifying patterns in the flow data is set out below in relation to Figure 2.

The flow monitor 110 may send some or all of the flow data, generated flow alerts and/or flow statistics generated from the flow data (e.g. data about flow events) to a remote datacentre, such as analysis server 180 on an external network, e.g. via its connection to the Internet 170. The flow monitor 110 may send the flow information to the analysis server continuously, e.g. as it is created, or periodically, e.g. every 30 mins, every hour, every 4 hours of 6 hours or every 12 or 24 hours. If the flow monitor 110 generates an alert, it may override the periodic sending of data and send the alert to the analysis server 180 immediately, or shortly after (e.g. within 1 or 2 minutes), it is generated.

The remote datacentre 180 can log the information received from the flow monitor 110 and store it for analysis. For example, the analysis server 180 (in addition to, or instead of the flow monitor 110) may identify patterns in the flow data which may be useful for suppressing flow alerts, e.g. as set out below in relation to Figure 2. The analysis server 180 may send to the flow monitor 110 details of identified patterns and/or instructions for suppressing flow alerts.

The flow monitor 110 may include a user interface, which may display flow information, including flow alerts and/or flow statistics. The flow monitor 110 may be configured to make an alarm sound in order to alert a user of a likely leak or unintentional flow. The user interface may allow a user to program the flow monitor 110 and switch off or disable the alarm.

The flow monitor 110 may also be configured to send information (e.g. alerts and/or flow statistics) to a user device 190 located outside the user’s premises on an external network and connected to the Internet 170. In some embodiments the flow monitor 110 may also be configured to send information to a further user device 192 located at the user premises, which can be connected wirelessly (or by wired connections) to the user’s local network, e.g. as shown the user device 192 can be in wireless communication with the flow 14 monitor 110. User devices 190, 192 may take the form of smartphones, tablet computers, personal computers, and the like. User devices may include an application for interacting with the flow monitor 110, e.g. for sending control commands to the flow monitor 110 and displaying information received from the flow monitor 110. Where the analysis server 180 is configured to generate flow alerts, these alerts may be provided to user device 190 or 192 by the analysis server 180, e.g. via the Internet 170.

In alternative embodiments one or both of the first temperature sensor 112 and the second temperature sensor 114 may be part of the flow monitor 110, for example they may be integrated in the flow monitor 110.

In the example of Figure 1, the flow monitor 110 uses temperature differences to determine flow in the conduit, however alternatives are possible. For example, flow in a conduit may be determined by a flow meter, such as a mechanical, magnetic or ultrasonic flow meter. E.g. a mechanical sensor may be placed in the conduit itself.

In some embodiments, the analysis server 180 may be configured to identify, based on flow data received from the flow monitor 110, flows that are likely to be unintentional or indicative of a leak, and then to generate flow alerts. The analysis server 180 may generate these alerts instead of or in addition to the flow monitor 110 generating the alerts. Where the flow monitor 110 does not generate alerts itself, the analysis server 180 can send the alerts to the flow monitor 110 for display on the user interface or for alerting the user via other means, e.g. by sounding an alert. The analysis server 180 may additionally or alternatively send alerts to the user devices 190, 192 to notify the user.

Where the flow monitor 110 is arranged to communicate with the analysis server periodically, e.g. according to a preset communication schedule, the flow monitor 180 may be configured to send flow alerts to the analysis server 180 to pass onto the remote user device 190 as they are generated (regardless of the timing of the communication schedule) to reduce the time taken to notify the user.

Although only a single flow monitor is described above, in some embodiments multiple flow monitors may be provided at different locations within the premises so that more information about fluid conduits can be collected. For example, there may be flow monitors for each of water, gas and oil conduits, where present. Alternatively, there may be multiple flow monitors on various different conduits of the same type, e.g. water pipes, or on different sections or branches of the same conduit. This may make it easier to ascertain the location of a leak. In some embodiments, a single flow monitor may be used to monitor multiple conduits (and/or multiple branches or locations of the same conduit), e.g. by receiving information from a plurality of first temperature sensors (and optionally a plurality of second temperature sensors, although sometimes a single temperature sensor measuring the ambient temperature sensor may suffice), each first temperature sensor (and where appropriate each second temperature sensor) being associated with a particular conduit (or section/branch of a conduit).

Figure 2 shows an exemplary method 200 for monitoring flow and providing alerts. The method 200 may be performed at a flow monitor 110 at a user premises or at a remote server 180. In some embodiments, some steps are performed at the flow monitor 110, whilst other steps are performed at the remote analysis server 180.

At step 210 a set of historical flow data is received. This historical flow data relates to flow detected in a fluid conduit over a first time period. The first time period generally extends over several days, and preferably at least two to three weeks. The duration of historic data may be selected such that it balances recent changes to a user’s routine with a sufficient amount of data to determine the difference between genuine patterns and noise. A favourable trade-off appears to be a duration of 2-3 weeks. A duration of less than 6 months, preferably less than 4 months or less than 3 months may be preferred. In order to identify patterns between different days of the week, longer periods (such as at least a month, or at least two months or at least three months) may be required. The flow data may comprise details of the times at which flow was detected during the first time period. A period of continuous flow may be classified as a “flow event”, and each flow event may have an associated start time and duration and/or end time. Flow events and their length or duration may be determined from the flow data. Alternatively or additionally, flow data may comprise an indication of flow in the conduit at each sampling time throughout the first time period. The sampling times could, for example, mirror the sampling rate of the temperature measurements, e.g. every 3 seconds, every 5 seconds, every 10 seconds, every 20 seconds or every minute. The indication of flow may be an indication of whether or not there was flow in the conduit at that time, and may be a binary flag (e.g. flow/no flow). Flow data can optionally include the rate of fluid flow.

At step 215 recurring flow patterns are determined from the set of historical flow data. The recurring flow patterns may identify periods during which fluid is generally used, or used for a long period of time. The historical flow data may be analysed to determine flow events of substantially continuous flow having start times and durations/end times as part of the process of determining flow events. In order to determine such recurring flow patterns, the data may be split into certain time intervals at a recurring time of day, for example hourly or half-hourly intervals. Patterns in consumption based on the recurring time intervals may be identified. For example, a probability or likelihood of flow occurring at each time interval throughout the day may be identified. In a typical household there could be a higher likelihood of flows in the morning (e.g. 6am to 8am) and the evening (e.g. 6p to 10pm), but lower likelihood during the rest of the day. The flow patterns generally take into account the time of day, but may further take into account the day of the week, and in particular may distinguish between weekdays and weekends. Further examples of determining recurring flow patterns are described below.

At step 220, alert suppression periods are designated based on the recurring flow patterns. A period may be designated as a suppression period if there is a high likelihood of (lengthy) flows occurring in that time period each day. Alert suppression periods may be designated by dividing the day into time slots, e.g. 30mins, 1 hour or 2 hours, and deciding for each time slot whether the likelihood of flows occurring during that time slot meets one or more suppression criteria. Further explanation of how alert suppression periods are selected can be found below.

In embodiments where alert suppression periods are designated remote from the flow monitor 110 (e.g. at the analysis server 180), alert suppression periods, details of these periods and/or a command to suppress alerts, or even disable flow monitoring, during the alert suppression periods may be sent to the flow monitor 110.

At step 225, subsequent flow data is received for a second time period. The second time period is after the first time period, although it need not follow the first time period directly. The subsequent flow data may be received in real time, e.g. as measurements at the conduit are made and flow data generated therefrom. As before, the subsequent flow data may include the times at which flow is detected, and optionally the rate of fluid flow. The flow data may comprise an indication of flow (e.g. flow or no flow) in the conduit at each sampling time, and optionally the rate at each sampling time. The flow data may include the start time and duration and/or end time for continuous flows.

At step 230 flow events are detected from the subsequent flow data to generate potential flow alerts. Each instance of flow, e.g. a period of continuous flow, may be classified as a “flow event”. In step 230, the flow data may be analysed to detect flow events that may be indicative of unusual or unintentional flow or a leak. In some embodiments, a flow event of continuous flow for longer than a certain threshold duration may be classed as indicative of unintentional flow and thus cause a potential flow alert to be generated. The threshold duration could be e.g. 15minutes, 20minutes or 25 minutes. In some embodiments, the flow must be at a substantially constant rate for at least the threshold duration in order to generate a potential flow alert.

In some embodiments the flow may be considered to be at a substantially constant rate if the flow rate does not change by more than a certain rate change proportion of its starting value. For example, if the flow rate does not change by more than 20% of its starting value.

At step 235 any potential flow alerts generated during the alert suppression periods are suppressed. This may involve preventing a user being informed about the alarm. For example, the alarm message will not be presented on the user interface of the flow monitor 110, and no alarm sound may be made. In embodiments where alert messages are sent to user devices 190, 192 such flow alert messages will not be generated or sent.

In some embodiments, monitoring of flow is disabled during alert suppression periods, e.g. the temperature sensors 112, 114 and/or the flow monitor 110 may be switched off, or in standby mode. Thus it may be possible to reduce power consumption. Alternatively, in embodiments where the flow monitor 110 transmits information (e.g. flow data and/or flow alerts) to the analysis server 180, the flow monitor may continue to record the flow data, but the communication with the analysis server 180 may be disabled in order to reduce power consumption. For example a wireless interface of the flow monitor 110 may be switched off or disabled during alert suppression periods.

However, it may be desirable to continue to monitor and record flow during alert suppression periods as this new data can be used to learn more about flow patterns or user behaviour and thus the recurring flow patterns identification may be improved (e.g. by basing on an increased amount of data) and/or adjusted to account for changes over time. Where flow pattern identification is performed by the analysis server 180, the communication between analysis server 180 and flow monitor 110 may still be disabled during alert suppression periods, while the flow monitor 110 continues to monitor the flow. The flow data for the alert suppression periods may then be transmitted subsequently, e.g. immediately following each alert suppression period or periodically, e.g. after a delay.

In some embodiments, details of the potential flow alerts may be recorded and stored, e.g. at the flow monitor 110 and/or at the analysis server 180.

Figure 3 shows an exemplary method 300 for identifying recurring flow patterns and designating alert suppression periods. This method 300 may be performed at steps 215 and 220 of the method 200 of Figure 2.

At step 310 the data is divided into time intervals or slots. For example, the data may be divided into time intervals of equal duration, e.g. 30mins, 1 hour, 2 hours or 3 hours.

At step 315 flow characteristics are attributed to each time interval. A flow characteristic may be binary, for example, an indication of whether or not flow occurred during the time interval, or may be a numeric quantity indicative of a characteristic of flow in the time interval (e.g. a number of distinct continuous flows, or the rate, or average rate, of flow in the conduit over the interval, or the total volume of fluid flow through the conduit in the time interval). Attributing characteristics to each time interval may comprise analysing the flow data to detect flow events. Where flow data is analysed to identify flow events, the indication of whether flow occurred during the time interval may be satisfied if any part of a flow event (e.g. start, middle, end) occurs during the interval. Alternatively, a flow may be considered to have occurred during the time interval only if the start (or only if the end) of a flow event occurs in that time interval.

In some cases, only continuous flows of more than a threshold flow duration are used to determine the flow characteristics for the time interval (e.g. short flows below the threshold duration may not be counted where the flow characteristic is the number of flows in the time interval, or where the flow characteristic is whether a flow has occurred, a time interval having only short flows be classified as not having any flow). The threshold flow duration could be e.g. 10 seconds, 20 seconds, 1 minute or 3 minutes. In some cases the threshold flow duration could be comparable (e.g. of the same order of magnitude) to the length of flow required to generate a potential flow alert, e.g. between 10 and 30 minutes, e.g. at least 15 minutes and/or not more than 25 minutes. The threshold flow duration for selecting time intervals for the recurring flow pattern is preferably less than or equal to the flow duration required to generate a potential flow alert.

In some cases, only flows of more than a threshold flow rate are used to determine the flow characteristics for the time interval (e.g. flows below the threshold rate may not be counted where the flow characteristic is the number of flows in the time interval, or where the flow characteristic is whether a flow has occurred, a time interval having only low rate of flow may be classified as not having any flow). The threshold flow rate may be e.g. around 1 litre per hour. In some embodiments the flow rate threshold is at least 0.1 litres/hour, at least 0.2 litres/hour or at least 0.5 litres/hour. Preferably the flow rate threshold is alternatively or additionally not more than 10 litres/hour or not more than 5 litres/hour.

In some embodiments both the threshold flow duration and threshold flow rate are used to determine characteristics (e.g. flow rate and flow duration must be above the respective thresholds for the flow to contribute to characteristics). Continuous long flows (e.g. more than 10 or more than 20 minutes) of very low flow rate may be less likely to be intentional flow.

The characteristics may take numerical values, for example the number of flow events that occur during that interval, or the proportion of time in that interval for which flow occurs. The flow characteristics may also include information about the flow rate (e.g. average flow rate during the interval).

At step 320 the time intervals are grouped according to periodic recurrence, e.g. a group of time intervals that repeat on a periodic basis. The recurrence may be daily or weekly. For example, time intervals occurring at the same time each day may be grouped together, e.g. having a 24 hour recurrence (e.g. the time interval 9am-10am for each day will be placed in the same group for analysis). In some situations, the time intervals at the same time each day will not be placed in the same grouping. For example, a distinction may be made between weekdays and weekends (e.g. the time intervals 9am-10am on MondaysFridays may be placed in one group, whilst time intervals 9am-10am on Saturdays and Sundays may be placed in a different group).

The time intervals need not be of equal duration. For example, intervals at times of day that tend to have less flow (e.g. night time) may be longer. Where time intervals are of unequal duration it is desirable that recurring instances of the time periods are divided in the same way e.g. so that time periods on one day correspond exactly to time periods on subsequent days.

At step 325 it is determined, for each group, whether a first flow criterion is satisfied. The first flow criterion is based on the time intervals in that group only, independent of time 20 intervals in other groups. The flow characteristics attributed to the time intervals at step 315 are used for this determination.

In one example, the first flow criterion is based on the proportion of time intervals in the group during which flow has occurred, e.g. the first flow criterion may be satisfied if a flow has occurred during at least a first threshold proportion of time intervals in that group. The first threshold proportion may be, e.g. around 2%, 5%, 10% or 15%. It is desirable to set the first threshold to a value which is large enough so as not to classify too many intervals as suppression periods (and therefore suggest unnecessarily long suppression periods), but the first flow threshold should also be set low enough that suppression periods are still detected where appropriate. The inventors have found that setting the first flow threshold to 5% represents a reasonable trade-off between the two cases. This first flow criterion may help to prevent a time interval being classified as an alert suppression period in the case that very few flows have been detected, but they happened to occur at the same time of day.

In another example, the first flow criterion may be satisfied if the average duration of flow in the time intervals in that group is above a first threshold flow value, or if the volume of flow in the time intervals in that group is above a first threshold volume.

In some embodiments, a plurality of first flow criteria, which are based only on data for time intervals in the group, may be used.

At step 330 it is determined, for each group, whether a second flow criterion is satisfied. The second flow criterion is not independent of time intervals in other groups, instead it is based on how the flow characteristics of time intervals in the group compare to those of time intervals in other groups (generally, in relation to all time intervals in the historical data).

In one example, the second flow criterion is based on the proportion of the number of total flows across all time intervals that occur during time intervals in the group. The second flow criterion may be satisfied if the number of flows occurring during time intervals in the group as a proportion of all flows in the set of historical data exceeds a second threshold proportion. The second threshold value may be around 5%, 7%, 10% or 15%. Generally the second threshold value will be less than 30%, preferably less than 25% or less than 20%. The second threshold value may be chosen based on the size of the time intervals, the number of groups and/or the time period between recurring intervals. For example, where time intervals are each an hour, and groups contain time intervals that recur on a daily 21 (24-hour) basis, the probability of a flow occurring in any individual hour is approximately 4% in the case of randomly occurring flows or a constant leak. Therefore it is desirable to select a second flow threshold above the value of the probability of a flow occurring randomly. However the second threshold should not be so high that intervals are not selected for suppression of alerts to reduce the likelihood of false alarms. The inventors have found that setting the second flow threshold to 10% represents a reasonable trade-off between the two cases. This criterion may help prevent a customer having all periods of the day classified as alert suppression periods in the case of a constant leak.

In another example, the second flow criterion is based on the volume of fluid flowing through the conduit during time intervals in the group as a proportion of the total volume of fluid flowing through the conduit across all time intervals (e.g. time intervals in all groups). The volume of fluid may be estimated from the rate and duration of flow. The second flow criterion would be satisfied if the proportion of volume of flow across all time periods in the historical data occurring in time intervals in the group exceeds a second threshold value, which may be around 5%, 7%, 10% or 15%.

In some embodiments, a plurality of second flow criteria, which are based on data for intervals in the specific group as well as for intervals in some/all other groups, may be used.

At step 335, recurring time intervals are designated as alert suppression periods if they correspond to groups that satisfy both the first flow criterion and the second flow criterion.

Figure 4 shows a chart of historical flow. The historical flow data was collected for one particular premises over a period of 90 days. The flow data has been grouped into hour-long slots throughout the day, recurring daily.

The chart of Figure 4 illustrates the probability of a flow occurring at each hour of the day, independent of flow occurring at any other hour of the day. Thus this essentially shows the number of instances of each time period in a group during which flow occurs as a proportion of all time periods in that group. It can be seen that on nearly 13% of days at least one flow occurs between the hours of 6am and 7am. Note that the probability of flow occurring at any hour of day is independent of flow in other hours, and as such that probabilities do not sum to 1 over a 24 hour period.

Figure 5 shows another chart of historical flow data for the same 90 day period as that of Figure 4. However the chart of Figure 5 shows the probability of a flow occurring in each hour of the day, given that a flow will occur at some point. Thus this shows the number of instances in each hour-long time period having flow (i.e. all time periods in a single group during which there is flow) as a proportion of the total number of periods having flow occurring across all 90 days (i.e. all time periods in all groups).

For the example data in Figures 4 and 5, the hour from 6am to 7am exceeds both the first flow threshold and the second flow threshold, and as such would be classified as an alert suppression period. However, the hour from 7am to 8am exceeds the first flow threshold but not the second flow threshold, and as such would not be classified as an alert suppression period. No other hours exceed either threshold, and therefore would also not be classified as an alert suppression period.

Figure 6 illustrates a hardware architecture of an exemplary flow monitor 600 which is arranged to monitor flow in a fluid conduit, such as water pipe, and may be used as the flow monitor 110 of Figure 1. The flow monitor 600 has a processor 602 and a volatile /random access memory 604 for storing temporary data and software code being executed. Random access memory 604 may be used to store, for example, flow data and/or measurements of characteristics indicative of flow in the fluid conduit as they are received or otherwise determined (e.g. by calculation), for example historical flow data and subsequent flow data as it is received. Random access memory 604 may also store details of identified flow patterns and/or alert suppression periods when they are determined (e.g. when calculated at the flow monitor) or received (e.g. from a remote server).

Flow monitor 600 also has persistent storage 606 which stores alert condition data 624, historical flow data 626 and alert suppression data 628. The alert condition data 624 may include conditions or logic for a potential flow alert to be generated, for example flow thresholds (e.g. flow durations) which would trigger an alert. The historical flow data 626 may include details of fluid flow in a fluid conduit for a first time period. The alert suppression data 628 may include details of alert suppression periods, for example, start and end times.

Persistent storage 606 may include other software and data, such as an operating system, device drivers, software configuration data, and the like.

The flow monitor 600 has a user interface 640 for presenting information, such as notifications or alerts, to a user and for receiving user inputs. This may be e.g. a screen and keyboard or a touch screen. The user interface 640 may also include a speaker, e.g. for sounding an alert to a user. Thus the user may be able to input settings for flow alerts (e.g. flow durations that will trigger an alert).

The flow monitor 600 also has a network interface 612 for communicating with devices at the premises and with remote devices. For example, network interface 612 may be comprise a wireless interface for connecting via Wi-Fi and/or Zigbee, e.g. to connect to user devices at the premises, and for connection to the Internet, such as for communication with a remote server and/or remote user devices. Network interface 612 may also include a wired interface for connection to sensors which monitor characteristics indicative of flow in the fluid conduit (e.g. a flow meter or temperature sensors). In alternative embodiments, the flow monitor may be configured to connect wirelessly to sensors which monitor characteristics indicative of flow. In other embodiments, the sensors to monitor characteristics indicative of flow may be integrated into the flow monitor 600.

The device components are interconnected by a data bus (this may in practice consist of several distinct buses such as a memory bus and I/O bus).

While a specific architecture is shown, any appropriate hardware/software architecture may be employed. For example, external communication may be via a wired network connection.

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.

Aspects of the apparatus and methods described herein are further exemplified in the following numbered CLAUSES:

1. A method for filtering alerts in response to flow in a fluid conduit, wherein alerts are triggered in response to detection of flow, the method comprising the steps of: receiving a set of historical flow data relating to flow in the fluid conduit for a first time period;

identifying one or more recurring flow patterns based on the set of historical flow data;

designating one or more alert suppression periods based on the identified recurring flow patterns;

receiving subsequent flow data relating to flow in the fluid conduit for a second time period;

detecting flow events in the subsequent flow data to generate one or more potential flow alerts; and suppressing potential flow alerts for flow events during the designated alert suppression periods within the second time period.

2. A method according to CLAUSE 1, wherein:

identifying a recurring flow pattern comprises:

dividing the first time period into time intervals; and grouping the time intervals into a plurality of groups of temporally recurring time intervals; and wherein designating one or more alert suppression periods is based on:

a first measure for each group that is independent of flow data of time intervals in the remainder of the plurality of groups; and/or a second measure for each group that is dependent on flow data for time intervals in at least two of the plurality groups.

3. A method according to CLAUSE 2, further comprising, after dividing the first time period into time intervals, obtaining flow characteristic data for each time interval from the historical flow data.

4. A method according to CLAUSE 3, wherein the flow characteristic data comprises one or more of:

an indication of whether or not there was a flow in the fluid conduit in the time interval;

a duration of flow in the fluid conduit during the time interval;

a rate or average rate of flow in the fluid conduit during the time interval;

a number of flow events detected in the time interval; and a flow volume of flow passing through the conduit in the time interval.

5. A method according to any of CLAUSES 2 to 4, wherein designating the one or more alert suppression periods comprises designating recurring time intervals corresponding to the groups of temporally recurring time intervals which satisfy both: a first criterion based on the first measure; and a second criterion based on the second measure.

6. A method according to any of CLAUSES 2 to 5, wherein the first measure is the proportion of time intervals in the group for which flow is detected.

7. A method according to CLAUSE 6 when dependent on CLAUSE 5, wherein the first criterion is satisfied if the proportion of time intervals in the group for which flow is detected is greater than a first threshold proportion.

8. A method according to CLAUSE 7, wherein the first threshold proportion is at least 1%, preferably at least 2% or 3%, more preferably around 5%.

9. A method according to CLAUSE 6 or 8, wherein the first threshold proportion is less than 30%, preferably less than 20%, more preferably less than 10%.

10. A method according to any of CLAUSES 7 to 9, wherein the first threshold proportion is at least 3% and not more than 8%.

11. A method according to any of CLAUSES 2 to 10, wherein the second measure is the proportion of total flows detected in time intervals in at least two groups that are detected in time intervals in the group.

12. A method according to CLAUSE 11 when dependent on CLAUSE 5, wherein the second criterion is satisfied if the proportion of flows in the at least two groups that are detected in time intervals in the group is greater than a second threshold proportion.

13. A method according to CLAUSE 12, wherein the second threshold proportion is at least 3%, preferably at least 4%, more preferably at least 8%, preferably around 10%.

14. A method according to CLAUSE 12 or 13, wherein the second threshold proportion is less than 40%, preferably less than 30%, more preferably less than 20% or 15%.

15. A method according to any of CLAUSES 12 to 14, wherein the second threshold proportion is at least 8% and not more than 15%.

16. A method according to any of CLAUSES 2 to 15, wherein the plurality of groups comprises at least two groups that are determined independently for different days of the week, for example being determined on weekdays independently of weekends.

17. A method according to any preceding CLAUSE, wherein suppressing potential flow alerts comprises changing one or more criteria for triggering an alert, such as increasing an alert threshold.

18. A method according to any preceding CLAUSE, wherein suppressing potential flow alerts comprises suppressing flow alerts entirely.

19. A method according to any preceding CLAUSE, further comprising:

updating the designation of the alert suppression periods using the subsequent flow data.

20. A method according to CLAUSE 19, wherein updating is performed periodically, e.g. at periodic time intervals such as every day, or every week.

21. A method according to any preceding CLAUSE, wherein the alert suppression periods and/or the time intervals are each at least 10 minutes, preferably at least 20 minutes, more preferably at least 30 minutes, or preferably around one hour.

22. A method according to any preceding CLAUSE, wherein the alert suppression periods and/or the time intervals are each less than around 4 hours, preferably less than three hours or two hours, more preferably around one hour.

23. A method according to any preceding CLAUSE, wherein the alert suppression periods and/or the time intervals are between 40 minutes and 90 minutes.

24. A method according to any preceding CLAUSE, wherein potential flow alerts are generated in response to the detection of fluid flowing in the fluid conduit continuously for a minimum flow time period.

25. A method according to CLAUSE 24, wherein potential flow alerts are generated in response to the detection of fluid flowing in the fluid conduit for the minimum flow time period at a rate between a lower flow rate threshold and an upper flow rate threshold.

26. A method according to CLAUSE 24 or 25, wherein the minimum flow time period is at least 2 minutes, preferably at least 5 minutes or 10 minutes, more preferably around 15 minutes.

27. A method according to any of CLAUSES 24 to 26, wherein the minimum flow time period is less than an hour, preferably less than 30 minutes or 20 minutes.

28. A method according to any of CLAUSES 24 to 27, wherein the minimum flow time period is between around 10 and 20 minutes.

29. A method according to any preceding CLAUSE, wherein:

the one or more potential flow alerts are each generated in response to detection of a flow event that is characteristic of unintentional flow.

30. A method according to any preceding CLAUSE wherein suppressing potential flow alerts comprises:

for each potential flow alert, establishing whether the potential flow alert has been detected during one of the one or more designated alert suppression periods; and upon establishing that the potential flow alert has been detected during one of the one or more designated alert suppression periods, preventing notification of a flow alert to a user.

31. A method according to any preceding CLAUSE wherein suppressing potential flow alerts comprises:

discarding or ignoring at least some or all of the received subsequent flow data relating to the one or more designated alert suppression periods prior to detecting flow events in the subsequent flow data to generate one or more potential flow alerts.

32. A method according to any preceding CLAUSE, wherein the set of historical flow data and/or the subsequent flow data comprises one or more of:

an indication of whether there is a flow in the fluid conduit;

a duration of a flow in the fluid conduit;

the temperature of the fluid conduit and/or the ambient temperature surrounding the fluid conduit; and a rate of flow in the fluid conduit.

33. A method according to any preceding CLAUSE, wherein the first time period is at least one week, preferably at least two weeks, more preferably around 3 weeks and/or preferably less than 2 months, or less than one month.

34. A method according to any preceding CLAUSE, further comprising alerting a user based on a potential flow alert generated outside the designated alert suppression periods.

35. A method according to CLAUSE 34, wherein the step of alerting a user comprises outputting a flow alert notification to a user interface; and/or sending a flow alert notification to a user device.

36. A method according to any preceding CLAUSE, wherein receiving a set of historical flow data relating to flow in the fluid conduit for a first time period and/or receiving subsequent flow data relating to flow in the fluid conduit for a second time period comprises making a measurement indicative of flow in the fluid conduit.

37. A method according to CLAUSE 36, wherein measuring flow in the fluid conduit comprises:

measuring a first temperature being the temperature of the outer surface of the fluid conduit;

measuring a second temperature being the ambient temperature outside of the fluid conduit; and making a flow determination by determining a measure of the convergence of the first and second temperatures over time.

38. A computer-readable medium comprising software code adapted, when executed on a data processing apparatus, to perform a method as set out in any preceding CLAUSE.

39. An apparatus for filtering alerts in response to flow in a fluid conduit, wherein alerts are triggered in response to detection of flow, the apparatus comprising:

a memory;

an interface for receiving flow data relating to flow in the fluid conduit; and a processor operable to:

receive a set of historical flow data relating to flow in the fluid conduit for a first time period;

identify one or more recurring flow patterns based on the set of historical flow data;

designate one or more alert suppression periods based on the identified recurring flow patterns;

receive subsequent flow data relating to flow in the fluid conduit for a second time period;

detect flow events in the subsequent flow data to generate one or more potential flow alerts; and suppress potential flow alerts for flow events during the designated alert suppression periods within the second time period.

40. A system for filtering alerts in response to flow in a fluid conduit, comprising:

an apparatus for filtering alerts in response to flow in a fluid conduit according to CLAUSE 39; and one or more sensors operable to make a measurement relating to flow in the fluid conduit.

41. A system for filtering alerts in response to flow in a fluid conduit according to CLAUSE 40, wherein the one or more sensors comprise one or more temperature sensors, preferably the one or more sensors comprising:

a first temperature sensor arranged to be mounted on the outer surface of the fluid conduit to measure a first temperature being the temperature of the outer surface and arranged to generate a first temperature signal indicative of the first temperature; and a second temperature sensor arranged to be positioned spaced apart from the first temperature sensor, to measure a second temperature being the ambient temperature outside of the fluid conduit and to generate a second temperature signal indicative of the second temperature.

42. A system for filtering alerts in response to flow in a fluid conduit according to CLAUSE 40 or 41, further comprising a fluid conduit.

43. An apparatus for filtering alerts in response to flow in a fluid conduit according to CLAUSE 39 or a system for filtering alerts in response to flow in a fluid conduit according to any of CLAUSES 40 to 42, the apparatus or system further comprising:

a user interface configured to present a flow alert to a user, such as a screen, for example a screen of a smartphone, personal computer, smart watch and the like, or a speaker.

Claims

1. A method for filtering alerts in response to flow in a fluid conduit, wherein alerts are triggered in response to detection of flow, the method comprising the steps of:

receiving a set of historical flow data relating to flow in the fluid conduit for a first time period;

2. A method according to claim 1, wherein:

identifying a recurring flow pattern comprises:

3. A method according to claim 2, further comprising, after dividing the first time period into time intervals, obtaining flow characteristic data for each time interval from the historical flow data.

4. A method according to claim 3, wherein the flow characteristic data comprises one or more of:

a duration of flow in the fluid conduit during the time interval;

a rate or average rate of flow in the fluid conduit during the time interval;

02 05 19 a number of flow events detected in the time interval; and a flow volume of flow passing through the conduit in the time interval.

5. A method according to any of claims 2 to 4, wherein designating the one or more alert suppression periods comprises designating recurring time intervals corresponding to the groups of temporally recurring time intervals which satisfy both: a first criterion based on the first measure; and a second criterion based on the second measure.

6. A method according to any of claims 2 to 5, wherein the first measure is the proportion of time intervals in the group for which flow is detected.

7. A method according to claim 6, wherein the first threshold proportion is at least 3% and not more than 8%.

8. A method according to any of claims 2 to 7, wherein the second measure is the proportion of total flows detected in time intervals in at least two groups that are detected in time intervals in the group.

9. A method according to claim 8 when dependent on claim 5, wherein the second criterion is satisfied if the proportion of flows in the at least two groups that are detected in time intervals in the group is greater than a second threshold proportion.

10. A method according to any of claims 2 to 9, wherein the plurality of groups comprises at least two groups that are determined independently for different days of the week, for example being determined on weekdays independently of weekends.

11. A method according to any preceding claim, wherein suppressing potential flow alerts comprises changing one or more criteria for triggering an alert, such as increasing an alert threshold, or wherein suppressing potential flow alerts comprises suppressing flow alerts entirely.

12. A method according to any preceding claim, further comprising:

13. A method according to any preceding claim, wherein the alert suppression periods and/or the time intervals are each at least 10 minutes.

14. A method according to any preceding claim, wherein the alert suppression periods and/or the time intervals are between 40 minutes and 90 minutes.

02 05 19

15. A method according to any preceding claim, wherein potential flow alerts are generated in response to the detection of fluid flowing in the fluid conduit continuously for a minimum flow time period.

16. A method according to claim 15, wherein the minimum flow time period is at least 2 minutes.

17. A method according to any preceding claim, wherein:

18. A method according to any preceding claim wherein suppressing potential flow alerts comprises:

for each potential flow alert, establishing whether the potential flow alert has been detected during one of the one or more designated alert suppression periods; and upon establishing that the potential flow alert has been detected during one of the one or more designated alert suppression periods, preventing notification of a flow alert to a user; or discarding or ignoring at least some or all of the received subsequent flow data relating to the one or more designated alert suppression periods prior to detecting flow events in the subsequent flow data to generate one or more potential flow alerts.

19. A method according to any preceding claim, further comprising alerting a user based on a potential flow alert generated outside the designated alert suppression periods.

20. A method according to claim 19, wherein the step of alerting a user comprises outputting a flow alert notification to a user interface; and/or sending a flow alert notification to a user device.

21. A method according to any preceding claim, wherein receiving a set of historical flow data relating to flow in the fluid conduit for a first time period and/or receiving subsequent flow data relating to flow in the fluid conduit for a second time period comprises making a measurement indicative of flow in the fluid conduit.

22. A computer-readable medium comprising software code adapted, when executed on a data processing apparatus, to perform a method as set out in any preceding claim.

23. An apparatus for filtering alerts in response to flow in a fluid conduit, wherein alerts are triggered in response to detection of flow, the apparatus comprising:

a memory;

02 05 19 an interface for receiving flow data relating to flow in the fluid conduit; and a processor operable to:

24. A system for filtering alerts in response to flow in a fluid conduit, comprising:

an apparatus for filtering alerts in response to flow in a fluid conduit according to claim 23; and one or more sensors operable to make a measurement relating to flow in the fluid conduit.

25. An apparatus for filtering alerts in response to flow in a fluid conduit according to claim

23 or a system for filtering alerts in response to flow in a fluid conduit according to claim

24 , the apparatus or system further comprising: