EP4600929A1

EP4600929A1 - Method for pre-alarm check arranged to detect an alarm prevention

Info

Publication number: EP4600929A1
Application number: EP24157093.6A
Authority: EP
Inventors: Arlete RODRIGUES; Andreas Kahl
Original assignee: Robert Bosch GmbH
Current assignee: Robert Bosch GmbH
Priority date: 2024-02-12
Filing date: 2024-02-12
Publication date: 2025-08-13

Abstract

Computer-implemented method for operating an alarm system having a plurality of signal detectors and an electronic data processor, said method using the electronic data processing, and comprising the steps of: receiving user input for modifying operation of the alarm system; applying the received user input to modify the operation of the alarm system; operating the alarm system for a predetermined first time period; recording pre-alarm and alarm data records during said first time period; receiving a plurality of pre-alarm and alarm data records each with a timestamp; displaying aggregated counts of said pre-alarm and alarm data records aggregated by a predetermined second time period.

Description

TECHNICAL FIELD

The present disclosure belongs to the technical field of smoke detectors, and more particularly to a computer-based method for operating an alarm system and detecting when errors occur in wiring and connections that may hinder the system from working correctly.

BACKGROUND

The impacts of false alarms on business, public and Fire Rescue Services (FRS) are enormous, such as:

Lost production time and associated costs;
Disruption of services provided by the affected business;
Associated costs to businesses of retained firefighters in their employment being released to respond;
Negative impact on the goodwill of employers who permit retained personnel to respond to automatic fire alarm activations which turn out to be false alarms;
A loss of confidence in the reliability of the premises fire alarm system, causing complacency amongst occupiers upon activation of the alarm which could potentially prejudice their safety;
The diversion of operational crews from real emergencies, potentially putting life and property at risk;
The potential for the safety of both firefighters and the public to be compromised whilst driving to a false alarm under emergency conditions;
Disruption to safety critical training being undertaken by FRS personnel;
The environmental impact of unnecessary appliance movements and drain on the public finances.

Document WO2014044675A1 discloses that in public or also private buildings, surveillance systems are often installed which have a large number of surveillance cameras, whereby the image data streams of the surveillance cameras are transmitted to a surveillance centre via a network. In order to improve availability, an evaluation device is provided for a surveillance system, the surveillance system having at least one system component for a surveillance area with at least one sensor for monitoring the surveillance area and/or for self-monitoring, with an input section for receiving sensor data S from the at least one sensor, having a data processing device for processing the sensor data S and for determining a processing result V, having an evaluation device for evaluating the processing result V and for generating a message M on the basis of the evaluation, having an output interface for outputting the message M, the message describing a reliability level of the system component.
Document EP 2 500 882 B1 relates to, both the alarm system and the fire and flammable gas alarm method, allow long-term and permanent monitoring of the output of the single detector, along with a strong CPU processing capacity, allowing the individual detectors to alarm early if their operating data is abnormal the alarm limit has been reached, so that an upstream alarm triggering is possible and an accident risk is nipped in the bud. In addition, in view of the detected evaluate historical data to see if an alarm is triggered and automatically assess if the detectors are OK, if the data sent is plausible, and if the detectors are in need of care or attention. All this significantly increases the safety factor of the alarm system.
These facts are disclosed in order to illustrate the technical problem addressed by the present disclosure.

GENERAL DESCRIPTION

Clearly, to encourage a reduction of the incidence of unwanted fire signals and the need for systems to provide early detection and warnings of a fire situation are imperative.
A key aspect of fire protection is to identify a developing fire emergency in a timely manner, and to alert the building's occupants and fire emergency organizations. This is the role of fire detection and alarm systems. These systems also self-monitor, identifying where within the building(s) alarms originate from and detecting when errors occur in wiring and connections that may hinder the system from working correctly.
In essence, a fire alarm system has four key functions: detect, alert, monitor and control. These sophisticated systems use a network of devices, appliances, and control panels to carry out these four functions.
Common fire detectors have, amongst other, optical, temperature and chemical sensors. Optical sensors use the scattered-light method. A LED transmits light to the measuring chamber, where it is absorbed by the labyrinth structure. In the event of a fire, smoke enters the measuring chamber, and the smoke particles scatter the light from the LED. The amount of light hitting the photo diode is converted into a proportional electrical signal.
The thermal sensor measures the temperature at regular intervals. Depending on the detector class, the temperature will trigger the status of alarm when temperature reaches values above 54°C or 69°C, or if the temperature increased in a specific value in a specific timeframe.
One of the main functions of the chemical sensor is to detect carbon monoxide generated as a result of a fire, but it also detects hydrogen and nitrous monoxide. The sensor signal value is proportional to the concentration of gas.
The way a fire alarm system detects a fire is through its initiating devices, which detect smoke or a fire. These devices include smoke detectors of various kinds, heat detectors of various kinds, sprinkler water flow sensors, and pull stations. The type of alarm and its settings should be established so as to enable the operator to make the necessary assessment and take the required timely action. Settings should be documented and controlled in accordance with the alarm system management controls.
A fundamental reason for having an alarm should be to prompt an operator action (i.e. making a control change). Frequently there are alarms implemented that require no operator action, e.g. simply for status indication. These alarms are "in" during normal operations. Although an argument can be made to justify the implementation of "monitoring" a value more closely (e.g. during a product switch in a processing system), this approach should be used with caution since it is an easy way to justify adding volumes of alarms.
There are also pre-alarms that gives a pre-warning at the main panel without operating the sounders, giving staff time to investigate a false alarm or potential fire. The detector still operates if it senses a fire. Pre-alarm provides an indication that the probability of an alarm is increased. Common causes may differ but usually are related with contamination of detector by ingress of dust, pollen, insects; steam from showers or aerosols sprays; an unsatisfactory maintenance and testing program in place; contractors working on site; accidental breakage of fire alarm call points; cooking fumes.
To mitigate false alarms and get customers satisfaction, the idea of the presented in this disclosure is to provide a service capable to identify recurrent changes on detectors behaviour and point the most critical systems/devices, providing timely info in a regular basis to the customers.
Despite advanced technology, false alarms do happen occasionally. This can be a result of faulty equipment, improper installation, user error, wrong sensitivity settings, wrong detector type, or any number of other issues. It is thus a need for an automatic process capable to alert for undesired situations that could put at risk the buildings and people's safety.
Data mining could help to explore fire alarm systems data in a regular basis, with low effort, identify problematic sources and work on solutions.
The process consists in getting the warnings (pre-alarms) and fire events reported by each available customer/system and identify the frequency of occurrence per month. This present disclosure has the advantage that the customer will be able to see the count of warnings/fires in a monthly basis triggered over in all the systems and can easily detect the most critical systems due the high number of pre-alarms. The present disclosure allows to timely act to identify the root cause and avoid system faults or false alarms, and with that avoid the inherent disturbances that lead to expensive costs with all the logistic, people management and business impacts.
The present disclosure discloses a computer-based method for operating an alarm system having a plurality of signal detectors and an electronic data processor, said method comprising the steps of: receiving user input for modifying operation of the alarm system; applying the received user input to modify the operation of the alarm system; operating the alarm system for a predetermined first time period; recording pre-alarm and alarm data records during said first time period; receiving a plurality of pre-alarm and alarm data records each with a timestamp; displaying aggregated counts of said pre-alarm and alarm data records aggregated by a predetermined second time period.
In an embodiment, the receiving user input comprises: receiving indication to ignore a pre-alarm, cancelling said pre-alarm and not triggering an alarm, or receiving indication to confirm a pre-alarm, not cancelling said pre-alarm and triggering an alarm.
In an embodiment, receiving user input further comprises: receiving indication to run a diagnostic on the alarm system and carrying out said diagnostic on the alarm system, and/or receiving indication to reset the alarm system and carrying out said reset of the alarm system.
In an embodiment, the electronic data processor is further arranged to process the pre-alarm data records with a pretrained machine-learning model to determine whether the pre-alarm record is determined to have an alarm data record, and the pre-alarm record is the alarm data record having to trigger an alarm; the pre-alarm record is the alarm data record having to cancel said pre-alarm and not trigger an alarm.
In an embodiment, the pretrained machine-learning model is an artificial recurrent neural network, a convolutional neural network, and/or a long short-term memory neural network.
In an embodiment, the electronic data processor is further arranged to determine if said the pre-alarm data records are periodical patterns over a longer period of time than said predetermined first time period.
In an embodiment, the electronic data processor is further arranged to input the pre-alarm and alarm data records to the pretrained machine-learning model by aggregating the pre-alarm and alarm data records for a subperiod of time comprised within said predetermined first time period.
In an embodiment, the predetermined first time period is from one week to one month.
In an embodiment, the predetermined second time period is from one month to one year.
In an embodiment, the signal detectors are a fire detector and/or a detector for flammable gas.
In an embodiment, the fire detector is a smoke detector or a thermal detector, or a smoke and temperature detector and the signal detector for flammable gas is a methane detector, a propane detector, or a carbon detector.
It is also disclosed an alarm system for detecting a fire or flammable gas released in an ambient, said system comprising: at least one signal detector installed in the ambient for detecting signals of the smoke, temperature, or flammable gas; an alarm control unit for real-time recording of the signals received by the signal detector; an electronic data processor to carry out the method of any of the previous embodiments.
It is also disclosed a computer program, configured to carry out every step of one of the methods of previous embodiments.
It is also disclosed a non-transitory machine-readable storage medium, on which the computer program of the previous embodiment is stored.

BRIEF DESCRIPTION OF THE DRAWINGS

The following figures provide preferred embodiments for illustrating the disclosure and should not be seen as limiting the scope of invention.

Figure 1 : Schematic representation of an embodiment of events list triggered by customer/system.
Figure 2 : Schematic representation of the number of Warnings/Fires alerts by month for customer X, system 1.
Figure 3 : Schematic representation of events details for a specific month (August 2021) selected in the warnings and fire statistics bar plot (figure 2).
Figure 4 : Schematic representation of an embodiment of the number of Warnings/Fires alerts by month for customer X (all systems), considering (top) and excluding (bottom) the walk test scenarios.

DETAILED DESCRIPTION

The present disclosure relates to a method for operating an alarm system and timely acting to identify the root cause and avoid system faults or false alarms.
It is expected that a system runs in a normal state, but if something different happens that affects the behaviour of any system's component, this event is recorded and stored in a Cloud database (Remote Alert). Through this database it is possible to have a clear identification of the system and device's properties that triggered the alarm (see example in Figure 1 ).
The data analysis process consists in getting the warnings and fire events reported in Remote Alert by each available customer/system and identify the frequency of occurrence per month. However, and to have a clear idea of what is happening in the fire alarm system and what could be the possible consequences, besides these initial statistics of number of warnings and fires, it is also possible to show a detailed list of all events triggered in a specific month. For that, the user only needs to select the required month, and this click option forwards to the list of all events in that month.
Through the Cloud implementation, the user will see these statistics through a bar plot for the fire and warnings count ( Figure 2 ) and, by clicking in a specific month bar a table will appear with all month's events details ( Figure 3 ).
To provide the most accurate results to the customer, there is a need to exclude the warnings coming from walk test scenarios. The needed tests required by the maintenance teams to evaluate the fire alarm system quality/stability affects the device's behaviour, leading them to reach the limits allowed for detecting a fire scenario. This process will trigger some trouble and warning events, which must not be considered in the analysis of potential false alarms.
A good example is shown in Figure 4 , where after removing all the warnings triggered after walk test scenarios, there is a considerable decrease in the number of warnings (very noticeable in August 2020).
In one embodiment of the disclosure, the algorithm is based on a simple count of events per month. From the Remote Alert data, the customer, system, state and timestamp are the properties that are required for this specific remote service.
Main steps:

extract event info from Remote Alert collection for the selected customer/system pair;
remove WARNING related with WALKTEST scenarios -> time interval to be considered: 3-min;
filter data by WARNING and FIRE events;
group the filtered data by timestamp (based on year/month);
separate the grouped data into FIRE and WARNING data frames;
elaborate the stacked bar plot based on the WARNING and FIRE count by month;
There are check boxes above the graph for Fire and Warning to allow the user filter out the graph;
include click bar to forward to a list of all events reported in that month;
- group data from step 1 by timestamp (based on year/month), not considering WARNING and FIRE events;
- join data frames from step 7.a. with data frames from step 6;
- Elaborate a table with all events details, count by month.

The service provided by this solution focus in three main levels:

i) Showing fire/warnings statistics to infer about the risks related with number of pre-alarms triggered;
ii) Comparing different systems at the same time and identify the most usual root causes for pre-alarms;
iii) Comparing different customers and identify the technical/installation issues more related with pre-alarms and fire alerts.

Public marketing, trade fairs, customer and market feedbacks can help in the provability on competitor product topic.
This disclosure will be used on the fire alarm systems helping to identify potential risks for alarms by look to pre-alarms.
The term "comprising" whenever used in this document is intended to indicate the presence of stated features, integers, steps, components, but not to preclude the presence or addition of one or more other features, integers, steps, components or groups thereof.
The disclosure should not be seen in any way restricted to the embodiments described and a person with ordinary skill in the art will foresee many possibilities to modifications thereof. The above-described embodiments are combinable. The following claims further set out particular embodiments of the disclosure.

Claims

Computer-implemented method for operating an alarm system having a plurality of signal detectors and an electronic data processor, said method using the electronic data processing, and comprising the steps of:
receiving user input for modifying operation of the alarm system;

applying the received user input to modify the operation of the alarm system;

operating the alarm system for a predetermined first time period;

recording pre-alarm and alarm data records during said first time period;

receiving a plurality of pre-alarm and alarm data records each with a timestamp;

displaying aggregated counts of said pre-alarm and alarm data records aggregated by a predetermined second time period.
Method according to the previous claim, wherein receiving user input comprises:
receiving indication to ignore a pre-alarm, cancelling said pre-alarm and not triggering an alarm, or

receiving indication to confirm a pre-alarm, not cancelling said pre-alarm and triggering an alarm.
Method according to any of the previous claims, wherein receiving user input further comprises:
receiving indication to run a diagnostic on the alarm system and carrying out said diagnostic on the alarm system, and/or

receiving indication to reset the alarm system and carrying out said reset of the alarm system.
Method according to any of the previous claims, wherein the electronic data processor is further arranged to
process the pre-alarm data records with a pretrained machine-learning model to determine whether

the pre-alarm record is determined to have an alarm data record, and

the pre-alarm record is the alarm data record having to trigger an alarm;

the pre-alarm record is the alarm data record having to cancel said pre-alarm and not trigger an alarm.
Method according to the previous claim, wherein the pretrained machine-learning model is an artificial recurrent neural network, a convolutional neural network, and/or a long short-term memory neural network.
Method according to any of the previous claims, wherein the electronic data processor is further arranged to determine if said the pre-alarm data records are periodical patterns over a longer period of time than said predetermined first time period.
Method according to any of the previous claims, wherein the electronic data processor is further arranged to input the pre-alarm and alarm data records to the pretrained machine-learning model by aggregating the pre-alarm and alarm data records for a subperiod of time comprised within said predetermined first time period.
Method according to any of the previous claims, wherein the predetermined first time period is from one week to one month.
Method according to any of the previous claims, wherein the predetermined second time period is from one month to one year.
Method according to any of the previous claims, wherein the signal detectors are a fire detector and/or a detector for flammable gas.
Method according to the previous claim, wherein the fire detector is a smoke detector or a thermal detector, or
a smoke and temperature detector and the signal detector for flammable gas is a methane detector, a propane detector, or a carbon detector.
Alarm system having a plurality of signal detectors for detecting signals of smoke, temperature, and/or flammable gas and an electronic data processor for detecting a fire or flammable gas released in an ambient, said system comprising:
an electronic data processor arranged to carry out the method of any of the claims 1-11.
Computer program, configured to carry out every step of one of the method of claims 1-11.
Non-transitory machine-readable storage medium, on which the computer program of claim 13 is stored.