Method for monitoring a heating system

Abstract

The present invention relates to a method for monitoring a heating device (1), in particular in a household appliance (2), wherein - a target heating process with target parameters is specified, - a recording device records an actual heating process with actual parameters and transmits it to a computer device (5), - the computer system (5) determines a deviation of the actual parameters of the actual heating process from the target parameters of the target heating process, - the computer system (5) adjusts the actual heating process to the target heating process or issues a signal to a user if the detected or calculated deviation reaches or exceeds a predefined limit.

Description

[0001] The present invention relates to a method for monitoring a heating device. The invention also relates to a computer program product and a household appliance, in particular an automatic cooking device, comprising a heating device, a detection device, and a computer device for carrying out the method. The invention further relates to a system comprising a household appliance, in particular an automatic cooking device, comprising a heating device, a detection device, and a computer device for carrying out the method, which is arranged separately from the household appliance and communicatively connected thereto.

[0002] Modern household appliances, particularly modern cooking appliances, use a variety of components that are subject to varying levels of wear and tear and therefore may have varying service lives. Such household appliances also use heating devices with heating elements to heat water, for example. If a heating device fails, it is common practice for a corresponding signal to be sent or issued to the user so that they can initiate repairs. However, such a signal is only generated after the respective component has failed, for example after the heating device has failed, which can lead to significant consequential damage to the respective household appliance.If a gradual deterioration in the functional properties of individual components occurs, for example, a decrease in the heating performance of a heating device over time, this is usually not indicated to the user. However, the resulting deterioration in both performance and, for example, the quality of the food produced has a negative impact on customer satisfaction.

[0003] The present invention therefore addresses the problem of providing a method by means of which the customer satisfaction of a user of a household appliance can be maintained in the long term.

[0004] This problem is solved according to the invention by the subject matter of independent claim 1. Advantageous embodiments are the subject matter of the dependent claims.

[0005] The present invention is based on the general idea of ​​monitoring a heating device of a household appliance, for example an automatic cooking device, preferably permanently, but at least regularly, and, upon detection of intolerable deviations between an actual heating output and a predetermined target heating output, either automatically initiating adjustment steps or otherwise providing the user with assistance. In the method according to the invention for monitoring the heating device, a target heating process, i.e. at least one such target heating process with associated target parameters, is thus specified, and an actual heating process with associated actual parameters is detected or recorded by means of a detection device and transmitted to a computer device. The computer device compares the actual parameters of the actual heating process with the target parameters of the target heating process and calculates corresponding deviations.The computer device then adapts the actual heating process, i.e. its actual parameters, to the target heating process, i.e. to the target parameters, or issues a signal to a user if the detected or calculated deviation reaches or exceeds a predefined limit value which, in particular, no longer allows automatic adjustment. If, for example, the heating output of the heating device continuously decreases, the computer device can in principle offer two alternative countermeasures if the detected or calculated deviation reaches or exceeds the previously defined limit value. Firstly, it is possible for the computer device to adapt the actual heating process or its actual parameters to the target heating process, i.e. its target parameters, by, for example, increasing the heating output or the heating duration if the fluid to be heated no longer reaches a predetermined temperature.Alternatively, it is of course also conceivable that if the calculated deviation reaches or exceeds the predefined limit, the computer device sends a corresponding signal to the user and, for example, suggests a service offer, allowing the user to seek external assistance. The method according to the invention thus makes it possible for the first time to continuously monitor a heating device or its heating output and to automatically initiate countermeasures as soon as a deterioration is detected, before the heating device fails completely. This not only increases customer satisfaction but can also reliably prevent potentially expensive consequential damage to the household appliance.

[0006] In an advantageous development of the method according to the invention, actual parameters at different points of a heating chain, such as in particular a heating element, an embedding element for the heating element and / or a heating plate, are recorded during the actual heating process and transmitted to the computer device, wherein the computer device then compares the recorded actual parameters with stored and complementary target parameters and initiates corresponding countermeasures, such as adjusting the actual parameters or generating a maintenance signal. Of course, other elements, such as connections or contacts, which contribute to homogeneous heat distribution, right down to the actual usable surface of a pot base, can also be monitored. The more different components are integrated into the heating chain, the greater the heat transfer resistance.If, for example, the heat transfer of a component deteriorates or its heat transfer resistance increases, this limits the overall heat transfer, which is the weakest link in the chain. Through extremely precise and therefore close monitoring of the heating chain, the computer system can even provide indications of possible fault locations during maintenance.

[0007] In an advantageous development of the method according to the invention, actual temperature curves with actual heating times and corresponding actual temperatures are recorded and transmitted to the computer. The computer then calculates an actual heating output from the recorded actual heating times, actual temperatures, and a mass to be heated according to the following formula: Plst=(c⋅m⋅ΔT) / t with: P Ist Actual heating output c constant m mass to be heated ΔT temperature difference T Time

[0008] This actual heating output is then compared with a target heating output and, according to the method according to the invention, appropriate countermeasures are then initiated, if necessary, such as adjusting the heating output by, for example, adjusting the actual heating times or issuing a signal to a user. Actual temperature curves from a wide variety of operating modes can be evaluated and, based on heating times, conclusions can be drawn about the effective heating output or energy density reaching the mass to be heated. By monitoring the heating device or its heating output, it is particularly possible to detect at an early stage whether the heating output is changing. Heating curves from steam programs within recipes are particularly suitable for this purpose, as these operate without temperature control, i.e. with a constant heating output. Furthermore, these programs require a pot content, i.e.an amount of water to be heated is precisely defined.

[0009] In a further advantageous embodiment of the method according to the invention, the deviation between the actual heating process and the target heating process is recorded during each heating process or at least periodically and transmitted to the computer device. This allows for continuous / permanent monitoring and early detection of damage to the heating system. By detecting damage to the heating system early on, a total failure and the potentially high consequential costs associated with it can be avoided.

[0010] In an advantageous development of the method according to the invention, the deviation of the actual heating process from the target heating process is calculated or estimated by the computer device using a KNN algorithm. This offers the great advantage that the heating characteristics can be trained and evaluated using artificial intelligence. With a KNN algorithm (K - nearest neighbor), the algorithm is first trained with a set of normal data points, i.e. in this case the normal heating behavior of the heating device. During a test phase, the number of neighboring data points that lie within a predefined distance, usually the so-called Euclidean distance, is then determined for each data point to be tested. This number is referred to as K. If a number of neighboring data points is below a certain threshold, i.e. a predefined limit, the data point is considered to be normal.The value of K influences the algorithm's sensitivity to these anomalies, with a smaller value of K leading to greater sensitivity, while a larger value of K leads to less sensitivity. The KNN algorithm is a non-parametric learning classifier that uses the concept of proximity to make classifications or predictions about groupings of a single data point. With the KNN algorithm, anomalies can also be calculated using artificial intelligence, thus allowing a failure, particularly of a heating device, to be more accurately estimated. The use of artificial intelligence (AI) can thus also increase customer satisfaction, as the probability of failure can be better estimated, thus preventing a total failure.

[0011] Before starting the KNN application, the data is prepared to process missing values, normalize the temperature values, and convert the variables into a numerical format suitable for the KNN algorithm or another method.

[0012] For example, if you have a set of training data {(x1,y1), ..., (x n ,y n )} for n elements, where Xi=xi1,xi2,xi3,…,x″i, is the variable set like timestamp, temperature, cooking step, cooking recipe and maybe other sensor data like humidity, air pressure, ambient temperature and y i The data classification is e.g. normal, abnormal, etc. In addition, a test dataset similar to the training set is used to classify the data. The Euclidean distance is used DIY, to measure the distance between data points within a classification. The Manhattan distance can also be used here.

[0013] The following steps are carried out: Calculating the distance: First, the algorithm calculates the distance between the input vector of the test data point to be classified or predicted and the input vectors of all data points in the training dataset.

[0014] Selection of the k nearest neighbors: The algorithm selects the k data points with the shortest distances to the test data point. These k data points are the "k nearest neighbors" of the test data point.

[0015] Classification or regression: In classification problems, the label of the test data point is determined by a majority vote of the k nearest neighbors. This means that the class with the most votes among the k nearest neighbors is assigned to the test data point. In regression problems, the value of the test data point is calculated as the average of the values ​​of the k nearest neighbors.

[0016] The hyperparameter k is the number of neighbors used for classification or regression. It is important to choose k appropriately, as too small a value of k can lead to higher variance due to considering only a few neighbors, while too large a value of k can lead to higher bias due to considering distant neighbors that may not be representative.

[0017] Alternatively, there are of course other AI methods such as Support Vector Machines (SVM), Decision Trees or Deep Learning models, which are also very suitable.

[0018] In an alternative embodiment of the method according to the invention, the deviation of the actual heating process from the target heating process, i.e. specifically a deviation of the recorded actual parameters from the specified target parameters, is calculated using a neural network. In this case, a neural network is used for the inventive method for anomaly detection, in which the algorithm is first trained with a set of normal data points. In concrete terms, this means that the normal heating behavior of the heating device is trained. The neural network learns the underlying patterns in the data and builds a model that represents these patterns. During the test phase, the neural network uses its model to make a prediction for each data point to be tested. If this prediction for a data point deviates significantly from its actual value, this data point is considered abnormal.The algorithm's resilience to anomalies can be controlled by adjusting the neural network's architecture, for example, by adjusting the number of hidden layers, the number of neurons in each layer, and the activation functions used. The major advantage of such neural networks is that they can produce meaningful results for machine learning, especially for artificial intelligence, with large data sets and many data dimensions and influencing factors.

[0019] Anomaly detection with neural networks can be implemented in several ways. One method is the use of autoencoders, a special type of neural network designed for unsupervised learning. Autoencoders are often used for anomaly detection because they are able to create a compressed internal representation of the data and identify deviations from this representation as anomalies.

[0020] Anomaly detection process with autoencoders: Dataset: Suppose you have a dataset with N data points. Each data point is represented by a vector Xi=xi1,xi2,xi3,…,x″i represented by input variables such as timestamp, temperature, cooking step, cooking recipe and perhaps other sensor data such as humidity, air pressure, ambient temperature.

[0021] Autoencoder architecture: The autoencoder consists of two main parts: the encoder and the decoder.

[0022] The encoder is a function f(x) that converts the input vector x i to a compressed internal representation z i Mathematically expressed: z i = f(x i )

[0023] The decoder is a function g(z) that returns the compressed representation z i on the reconstructed output vector x′i Expressed mathematically: x′i=g(zi)

[0024] Training: The goal of training is to train the autoencoder so that it is able to convert the input vectors x i as accurately as possible. This is usually achieved by minimizing a reconstruction loss function that calculates the difference between the original input vectors x i and the reconstructed vectors x′i quantified. A common loss function is the mean squared error (MSE): L(xi,x′i)=‖xi−x′i‖2 where ∥....∥2 is the quadratic norm.

[0025] Anomaly detection: After training, the encoder is able to generate a compressed representation i of the input vectors. Anomalies can then be detected by calculating the reconstruction error for each data point. A larger reconstruction error than the average error observed in the training phase may indicate an anomaly.

[0026] Threshold: To interpret the reconstruction error and identify anomalies, a threshold must be set. Data points with a reconstruction error above the threshold are classified as anomalies.

[0027] The exact implementation of the autoencoder and the choice of loss function may depend on the specific application domain and the characteristics of the dataset. Autoencoders can be very powerful models for anomaly detection, especially when the data is high in dimensionality and may contain complex variations.

[0028] In a further advantageous embodiment of the method according to the invention, the computer device increases a heating duration and / or a heating output if the detected deviation reaches or exceeds the predefined limit value. In the event that the computer device is capable of adapting the actual heating process itself, it is thus possible to improve the heating, for example by increasing the heating duration and / or the heating output. In this case, the reduced heating output of the heating device can be compensated for internally without initially leading to a total failure and also without a user having to be specifically informed of this. In purely theoretical terms, of course, not only an increase in the heating output / heating duration is possible, but also a general adjustment of the actual parameters in all directions.

[0029] The present invention is further based on the general idea of ​​specifying a computer program product comprising instructions which, when executed by a household appliance or a computer device thereof, cause the latter to carry out the method described in the previous paragraphs. The method is thus preferably carried out by means of a correspondingly designed computer program product which is stored, in particular stored, on a non-volatile data storage device, in particular a memory device. The computer program product is advantageously stored in a part of the computer device or the household appliance. It is understood that the computer program product as such also falls within the scope of this invention.

[0030] The present invention is further based on the general idea of ​​providing a household appliance, in particular an automatic cooking appliance, with a heating device, a detection device, and a computing device. The computing device is designed to carry out the method described in the preceding paragraphs and is arranged in the household appliance itself. Both the detection and evaluation of the actual heating process take place in the household appliance itself, as does the adjustment of the actual heating process to the desired heating process or the output of a signal to the user.

[0031] The present invention is further based on the general idea of ​​specifying a system with a household appliance, in particular an automatic cooking appliance, having a heating device, a detection device, and a computing device. The computing device is designed to carry out the method described in the previous paragraphs and is arranged separately from the household appliance, and in this case is communicatively connected to the household appliance. According to the second alternative embodiment, the recorded actual heating processes or the recorded actual parameters of the actual heating process are transmitted externally to a backend and evaluated there. In this case, the externally arranged computing device (backend) triggers a corresponding reaction in the household appliance or initiates maintenance if required. The notification, for example, for maintenance, can also be sent to a connected smart device, such as a smartphone, Alexa, etc.The major advantage here is that the algorithms for evaluating the data in the external computer system can be continuously refined, and no software update is required on the household appliance to adapt the algorithms accordingly.

[0032] With the method according to the invention, the computer program product according to the invention, the household appliance according to the invention, and the system according to the invention, important insights can be gathered into how a heating device changes over its service life or over the service life of a household appliance. Furthermore, it is possible to improve customer service because a user or customer can be informed of this before the heating device actually fails, and maintenance of the household appliance can be initiated. Furthermore, the method according to the invention makes it possible to automatically adapt the parameters time and used heating power in corresponding operating sequences with fixed, predetermined process times and in processes that do not utilize the full rated power of the heating device.

[0033] Further important features and advantages of the invention emerge from the subclaims, from the drawings and from the associated description of the figures based on the drawings.

[0034] It is understood that the features mentioned above and those to be explained below can be used not only in the respective combination specified, but also in other combinations or on their own, without departing from the scope of the invention as defined by the claims. Components mentioned above and those to be mentioned below of a higher-level unit, such as a device, apparatus, or arrangement, which are designated separately, may form separate parts or components of this unit or be integral areas or sections of this unit, even if this is shown differently in the drawings.

[0035] Preferred embodiments of the invention are illustrated in the drawings and are explained in more detail in the following description, wherein the same reference numerals refer to the same or similar or functionally identical components.

[0036] They show, schematically, Fig. 1 individual method steps of a method according to the invention for monitoring a heating device, Fig. 2 a domestic appliance designed as a cooking machine with a heating device and a computer device for carrying out the method, Fig. 3 a system with such a household appliance with a heating device, a detection device and a computing device, wherein the computing device is arranged separately from the household appliance.

[0037] According to the Fig. 1 shows a method for monitoring a heating device 1 (see also Fig. 2) comprises several process steps A to D. The heating device 1 serves to heat, for example, a fluid, in particular water or a food, in a household appliance 2, which according to the Fig. 2 is designed, for example, as a cooking appliance 3. In method step A, a target heating process with target parameters is first specified and in method step B an actual heating process with actual parameters is recorded during operation of the household appliance 2 by means of a recording device 4 and transmitted to a computer device 5. In method step C, a deviation of the actual heating process or the actual parameters recorded thereby from the target heating process, i.e. from the specified target parameters, is determined. If the recorded deviation between the actual parameters and the target parameters reaches or exceeds a predefined limit value, the computer device 5 now adapts the actual heating process or its actual parameters to the target heating process or the target parameters in method step D or outputs a signal to a user.This makes it possible to monitor the heating device 1 and to detect a drop in performance at an early stage, whereby as a countermeasure either an adjustment of the actual heating process or its actual parameters is carried out or a signal is issued by means of which a user can, for example, inform a maintenance service.

[0038] In a further development of the method according to the invention, during the actual heating process, i.e. during normal operation, actual parameters at different points in a heating chain, in particular on a heating element, an embedding element for the heating element and / or a heating plate are recorded by means of the recording device 4 and transmitted to the computer device 5, wherein the computer device 5 compares the recorded actual parameters with stored and associated target parameters. This makes it possible to monitor not only an isolated component, but the heating chain per se, whereby an entire heat transfer chain from the heating element to a heating plate or a pot base can be monitored. The more different components are involved in the heat transfer chain, the greater the heat transfer resistance. If the heat transfer of a component orIf the heat transfer resistance of these increases, this limits the overall heat conduction in the sense of the weakest link in the chain.

[0039] Actual temperature curves with actual heating times and corresponding actual temperatures can be recorded as parameters and transmitted to the computer device 5. The computer device 5 calculates an actual heating output from the recorded actual heating times, the actual temperatures, and a mass to be heated according to the following formula: Plst=(c⋅m⋅ΔT) / t with: P Ist Actual heating output c constant m mass to be heated ΔT temperature difference T Time

[0040] This calculated actual heating output is now compared with a predefined target heating output, which corresponds to the heating output of a properly functioning heating device 1.

[0041] In order to enable the most seamless monitoring of the heating device 1 and in particular to avoid a total failure, the deviation of the actual heating process from the target heating process is preferably recorded during each heating process and transmitted to the computer device 5.

[0042] The deviation of the actual heating process from the target heating process can be calculated by the computer device 5, for example, using artificial intelligence and the use of ANN algorithms. Such ANN algorithms can detect anomalies relatively reliably. With such an ANN algorithm, the algorithm is first trained with a set of normal data points, which in the present case would occur during normal operation of the household appliance 2. During the test phase, for each data point to be tested, the number of neighboring data points is determined that lie within a certain distance, which usually corresponds to the Euclidean critical distance. This number is referred to as K. If the number of neighboring data points is below a certain threshold, the data point is considered anomalous.The sensitivity of the algorithm can be controlled by the value of K, with a small value of K resulting in higher sensitivity and a large value resulting in lower sensitivity.

[0043] Alternatively, it is also conceivable for the deviation of the actual heating process from the target heating process to be calculated by the computer device 5 using a neural network. Here, too, it is possible to use artificial intelligence to detect anomalies. When using a neural network, the algorithm is first trained with a set of normal data points. This set of normal data points simulates the normal operation of the household appliance 2. The neural network learns the underlying patterns in the data and builds a model that represents this pattern. During the testing phase, the neural network uses its model to make a prediction for each data point to be tested. If the prediction for a data point deviates significantly from its actual value, this data point is considered anomalous.The sensitivity of such a neural network to anomalies can be controlled by adjusting the architecture of the neural network, for example, the number of hidden layers, the number of neurons in each layer, and the activation functions used.

[0044] Regardless of the artificial intelligence method used, the method according to the invention can be used to gather important insights into how the heating device 1 changes over the entire service life of the household appliance 2. This makes it possible to improve customer service, since, for example, the customer can be informed even before the heating device 1 fails completely and maintenance of the household appliance 2 can be initiated. Furthermore, the method's ability to initiate countermeasures for a gradual reduction in performance of the heating device 1 can increase the performance of the household appliance 2 and thus also customer satisfaction.

[0045] Looking at the Fig. 3, one can recognize a system 6 with such a household appliance 2, which can be designed, for example, as a cooking appliance 3 and can in turn have a heating device 1, a detection device 4 and a computer device 5. In this case, however, the computer device 5 can also be arranged externally to the household appliance 2 and be communicatively connected thereto. The computer device 5 can be in the Fig. 3, a signal can be sent, for example, to a user's smartphone 7, if a detected deviation of the actual heating process from the target heating process reaches or exceeds a predefined limit. Data, such as weighing data, temperature curves, etc., can be exchanged between the computer device 5 and the household appliance 2, wherein the signal to the user for contacting a service provider is also transmitted from the computer device 5 to the household appliance 2 and, for example, via a display device 8 (see Fig. 2) or a display can be output.

[0046] The method according to the invention is preferably implemented using a computer program product comprising instructions that, when executed by the household appliance 2 or a computer device 5 thereof, cause the computer device 2 or these devices to execute the method. The computer program product can be stored in a non-volatile memory in the household appliance 2 or in a cloud.

[0047] With the method according to the invention, the computer program product according to the invention, the household appliance 2, and the system 6 according to the invention, it is possible to monitor a heating device 1 of such a household appliance 2, preferably continuously, and to quickly and reliably detect underperformance or gradually occurring errors and, if necessary, take countermeasures. This can, in particular, increase customer satisfaction and reliably avoid consequential costs that may arise in the event of a total failure of the heating device 1. Reference symbol 1 heating device 2 household appliances 3 Cooking machine 4 Recording device 5 Computer setup 6 Systems 7 smartphones 8 Display device / display

Claims

[1] Method for monitoring a heating device (1), in particular in a household appliance (2), in which - a target heating process with target parameters is specified, - an actual heating process with actual parameters is recorded by a recording device (4) and transmitted to a computer device (5), - the computer device (5) determines a deviation between the recorded actual parameters of the actual heating process and the target parameters of the target heating process, - the computer device (5) adapts the actual heating process to the target heating process or outputs a signal to a user if the detected deviation reaches or exceeds a predefined limit value. [2] Method according to claim 1, characterized bythat during the actual heating process, actual parameters are recorded at different points of a heating chain, in particular a heating element, an embedding element for the heating element and / or a heating plate, and transmitted to the computer device (5), wherein the computer device (5) compares the recorded actual parameters with stored and complementary target parameters. [3] Method according to claim 2, characterized by that actual temperature curves with actual heating times and associated actual temperatures are recorded as actual parameters and transmitted to the computer device (5). [4] Method according to claim 3, characterized by that the computer device (5) calculates an actual heating power P from the recorded actual heating times, actual temperatures and a mass to be heated Ist calculated according to the following formula Plst=(c⋅m⋅ΔT) / t with: P Ist Actual heating output c constant m mass to be heated ΔT temperature difference T Time and with a target heating output P SOLL compares. [5] Method according to one of the preceding claims, characterized by , that the deviation of the actual heating process from the target heating process is recorded by the recording device during each heating process and transmitted to the computer device (5). [6] Method according to one of the preceding claims, characterized by that the deviation of the actual heating process from the target heating process is calculated by the computer device (5) by means of a KNN algorithm, or that the deviation of the actual heating process from the target heating process is calculated by the computer device (5) by means of a neural network. [7] Method according to one of the preceding claims, characterized bythat the computer device (5) adjusts a heating time and / or a heating power if the detected deviation reaches or exceeds the predefined limit value. [8] Computer program product, comprising instructions which, when the computer program product is executed by the computer device (5), cause the computer device (5) to carry out the method according to one of claims 1 to 7. [9] Household appliance (2), in particular a cooking machine (3), with a heating device (1), a detection device (4) and a computer device (5), wherein the computer device (5) is designed to carry out the method according to one of the preceding claims. [10] System (6) with a domestic appliance (2), in particular a cooking machine (3), with a heating device (1), a detection device (4) and a computer device (5), wherein the computer device (5) is designed to carry out the method according to one of the preceding claims and wherein the computer device (5) is arranged separately from the domestic appliance (2) and is communicatively connected thereto.

Images