A real-time state evaluation and fault prediction method and system for a deicing system

Abstract

The application discloses a kind of deicing system real-time state evaluation and fault prediction method and system, the method includes: through the ground state monitoring test of deicing system simulates deicing system fault state, obtains the sensor data under different fault states;Different fault states are used to train state evaluation model and fault prediction model by the sensor data set constructed with data;Sensor is installed on the aircraft deployed with deicing system, to collect the working data of deicing system;The data collected by sensor is input into trained state evaluation model and fault prediction model, the real-time state of deicing system is evaluated, and the possible fault is predicted.The application can evaluate current anti-icing system working state according to current state monitoring data, and predict the possible fault of system.

Description

Technical Field

[0001] This invention belongs to the field of aircraft de-icing technology, and particularly relates to a method and system for real-time status assessment and fault prediction of a de-icing system. Background Technology

[0002] In the field of aerospace technology, icing on aircraft surfaces poses serious safety hazards. The patent applicant previously proposed a de-icing device, system and method for aircraft (CN114655443B), which mainly uses electric heating to reduce the adhesion between the ice layer and the wing skin. Then, a vibration is generated by an impact force generator based on electric pulse vibration, causing the surface ice to fall off. This de-icing system is particularly suitable for UAVs with limited onboard energy.

[0003] As drones play an increasingly prominent role in the aviation field, they have been widely used. However, when drones operate for extended periods in various complex weather conditions, the safe and stable operation of the anti-icing and de-icing system, especially timely understanding of the current working status of the de-icing system and possible malfunctions or performance degradation during operation, is crucial to the flight safety of drones.

[0004] Chinese patent application CN201911316405.5 discloses a method and system for detecting faults in an anti-icing and de-icing system based on timing judgment. The method includes: pre-setting an input timing signal according to the divided aircraft de-icing zones; controlling a switch array to activate the corresponding anti-icing and de-icing execution components of the anti-icing and de-icing system according to the input timing signal; collecting the working status information of the corresponding anti-icing and de-icing execution components; comparing the collected working status information with preset working status values ​​to generate corresponding logical status information; performing logical operations on the generated logical status information and the input timing signal to generate an output timing signal; calculating the time difference between the corresponding states of the input timing signal and the output timing signal; and determining whether a system fault has occurred by comparing the time difference with a preset time difference value.

[0005] The aforementioned existing technology uses "preset input timing signals," "preset operating state values," and "preset time difference values" for fault diagnosis. Its core is a timing comparison method based on fixed logic rules and manually set thresholds. This method requires pre-defining the timing logic and parameter boundaries for normal system operation. When the de-icing system deviates from the preset model due to aging, environmental changes, or complex faults, it is difficult to accurately identify unexpected faults or performance degradation states. Summary of the Invention

[0006] The purpose of this invention is to provide a method and system for real-time status assessment and fault prediction of a de-icing system, which partially solves or alleviates the above-mentioned deficiencies in the prior art. It can assess the current working status of the de-icing system based on current status monitoring data and predict possible faults in the system.

[0007] To solve the aforementioned technical problems, the present invention specifically adopts the following technical solution:

[0008] A first aspect of the present invention is to provide a method for real-time status assessment and fault prediction of a de-icing system, comprising:

[0009] A ground condition monitoring test of the de-icing system simulates the normal operation, open circuit, and fault conditions of the de-icing system, acquiring sensor data under each condition. The sensor data includes: temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the accelerometer under normal operation of the electric heating device and vibration device in each cycle; temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the accelerometer under fault conditions of the electric heating device and vibration device in each cycle; and temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the accelerometer under open circuit conditions of the electric heating device / vibration device in each cycle.

[0010] Sensor data under normal, open-circuit, and fault states are preprocessed to extract a training dataset, which is then used to train a state assessment model and / or a fault prediction model. Specifically, the preprocessing includes extracting the time-domain and frequency-domain features of the sensor data collected in each cycle, and constructing a 2×n state matrix as state labels, with each state matrix corresponding to a set of time-domain and frequency-domain features; where n is the number of partitions of the electric heating device or vibration device; in the row representing the state of the electric heating device, each element represents an electric heating partition, and the value of the element represents the state of the electric heating device; in the row representing the state of the vibration device, each element represents a vibration partition, and the value of the element represents the state of the vibration device.

[0011] Sensors are installed on aircraft equipped with de-icing systems to collect operational data from the de-icing systems.

[0012] The data transmitted back from the sensors is input into the trained state assessment model and fault prediction model to assess the real-time state of the de-icing system and predict possible faults.

[0013] Furthermore, the operating data of the electric heating device in the de-icing system includes temperature and current; the operating data of the vibration device in the de-icing system includes vibration frequency, vibration amplitude, and current.

[0014] Furthermore, for the electric heating device in the de-icing system, the temperature of the skin is monitored by a temperature sensor installed inside the skin, and the current is monitored by a current sensor.

[0015] For the vibration device in the de-icing system, the vibration frequency and vibration amplitude are monitored by an acceleration sensor installed on the skin near the vibration exciter or at the exciter support, and the current is monitored by a current sensor.

[0016] Furthermore, the de-icing system includes:

[0017] Multiple de-icing zones are symmetrically arranged along the fuselage axis; each de-icing zone is equipped with an independent heating device and a vibration device; the heating device is used to melt the ice layer and reduce the adhesion between the ice layer and the skin, and the vibration device is used to generate pulse vibrations on the skin to cause the ice layer on the skin to fall off; wherein, the multiple de-icing zones are pre-defined according to the maximum de-icing power provided by the power supply system.

[0018] The partition control module is used to set a rotation cycle and perform partition rotation control on multiple de-icing partitions based on the rotation cycle, so that the multiple de-icing partitions take turns to perform de-icing work, and the de-icing partitions performing de-icing simultaneously are symmetrical along the fuselage axis; each rotation cycle includes a number of partition working times; the number of partition working times is at least 1 / 2 of the number of de-icing partitions, so that at least two de-icing partitions symmetrically arranged along the fuselage axis work simultaneously during each partition working time; there are intervals between the partition working times;

[0019] The working time of each de-icing zone includes the working time of the heating device, the delay time, and the working time of the vibration device, which are set sequentially. Preferably, on the left and right wings symmetrical about the fuselage axis, each de-icing zone works in a cycle sequence; at any given time (i.e., within one zone's working time), there can be at most two de-icing zones on the left and right wings, while for a particular wing, there can be at most one de-icing zone in operation.

[0020] Furthermore, the steps for defining de-icing zones according to the maximum de-icing power provided by the power supply system include:

[0021] The maximum area of ​​a single de-icing zone is calculated using the formula s≤Q / (q*n), where s is the maximum area of ​​the de-icing zone, Q is the maximum de-icing power, q is the power of the electric heating film per unit area, and n is the number of de-icing zones working simultaneously; where q satisfies the requirement that the skin reaches the specified temperature within a preset time.

[0022] Furthermore, when constructing the dataset, a 2×n state matrix is ​​constructed as a sample, where n is the number of partitions of the electric heating device or the vibration device. In the row representing the state of the electric heating device, each element represents one partition of electric heating, with a value of 0 indicating that the electric heating device is open-circuited, a value of 1 indicating that the electric heating device is normal, and a value of 2 indicating that the electric heating device is abnormal. In the row representing the state of the vibration device, each element represents one partition of vibration, with a value of 0 indicating that the vibration system is open-circuited, a value of 1 indicating that the vibration device is normal, and a value of 2 indicating that the vibration device is abnormal.

[0023] Furthermore, the step of identifying the de-icing system's cycle includes:

[0024] Collect temperature sensor data for a single zone of the de-icing system over multiple cycle periods;

[0025] The starting point of the electric heating device in each cycle is selected from the temperature sensor data to obtain the variable point data point sequence;

[0026] The polling cycle is calculated based on the time interval between adjacent variable data points.

[0027] Furthermore, using the formula:

[0028] ;

[0029] Calculate the cycle period; where T i For the i-th cycle, D i Let D be the index value of the i-th working starting point. i+1 is the index value of the (i+1)th working starting point, and m is the sampling frequency.

[0030] Furthermore, before training the state assessment model and the fault prediction model, the dataset is preprocessed; the preprocessing includes data cleaning, feature extraction, and data normalization.

[0031] The present invention also provides a real-time status assessment and fault prediction system for a de-icing system, comprising:

[0032] The fault status data acquisition module is used to simulate the fault status of the de-icing system through ground status monitoring tests of the de-icing system, and to acquire sensor data under different fault statuses. The sensor data includes: temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the acceleration sensor under normal operating conditions of the electric heating device and vibration device in each cycle; temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the acceleration sensor under fault status of the electric heating device and vibration device in each cycle; and temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the acceleration sensor under open-circuit status of the electric heating device / vibration device in each cycle.

[0033] The model training module is used to preprocess sensor data under normal, open circuit, and fault conditions to obtain a training dataset, and to train the state assessment model and fault prediction model using the dataset.

[0034] The working data acquisition module is used to install sensors on aircraft equipped with de-icing systems to collect working data of the de-icing system;

[0035] The evaluation and prediction module is used to input the data returned by the sensors into the trained state evaluation model and fault prediction model to evaluate the real-time state of the de-icing system and predict possible faults.

[0036] Furthermore, the de-icing system includes:

[0037] Multiple de-icing zones are symmetrically arranged along the fuselage axis. Each de-icing zone is equipped with an independent heating device and a vibration device. The heating device is used to melt the ice layer and reduce the adhesion between the ice layer and the skin. The vibration device is used to generate pulse vibrations on the skin to cause the ice layer on the skin to fall off. The multiple de-icing zones are determined in advance according to the maximum de-icing power provided by the power supply system.

[0038] The partition control module is used to set the rotation cycle and perform partition rotation control on multiple de-icing partitions based on the rotation cycle, so that multiple de-icing partitions take turns to perform de-icing work, and the de-icing partitions performing de-icing at the same time are symmetrical along the fuselage axis.

[0039] Each cycle includes several partition working times; the number of partition working times is at least 1 / 2 of the number of de-icing partitions, such that at least two de-icing partitions symmetrically arranged along the fuselage axis work simultaneously during each partition working time; there are intervals between the partition working times.

[0040] The working time of the partition includes the working time of the heating device, the delay time, and the working time of the vibration device, which are set sequentially.

[0041] Beneficial effects:

[0042] 1. Data-Driven Innovative Assessment Methods. This approach abandons traditional state assessment models based on prior physical knowledge and relies entirely on data-driven methods. By collecting extensive sensor data from the de-icing system under different operating conditions, a labeled dataset is constructed, and machine learning methods are used to build state assessment and fault prediction models. This method is not limited by a deep understanding of the system's physical principles, can adapt to complex and changing actual operating conditions, and opens up new avenues for state assessment of aircraft thermo-coupled de-icing systems.

[0043] 2. Real-time and accurate status assessment. By acquiring real-time data from multiple sensors across various zones of the de-icing system, including temperature, current, and vibration frequency, and inputting this data into a trained status assessment model, the system's current real-time status can be quickly and accurately obtained. A 2×n status matrix visually presents the operating status (normal, open circuit, or abnormal) of each zone of the electric heating and vibration systems, helping operators to grasp the system's operational status immediately and promptly identify potential problems.

[0044] 3. Fault prediction ensures flight safety. A fault prediction model trained based on historical multi-sensor data can effectively predict potential anomalies / failures in the de-icing system. By analyzing trends and characteristic changes in the data, it can detect signs of system performance degradation in advance, such as aging of the electric heating film or abnormalities in vibration device components. Maintenance recommendations are provided before a failure occurs, giving ground personnel ample time for predictive maintenance or component replacement. This significantly reduces the risk of flight accidents caused by de-icing system failures, effectively ensuring UAV flight safety.

[0045] 4. Comprehensive Experimental Data Supports the Model. To ensure the accuracy and reliability of the model, this approach involves conducting ground-based condition monitoring experiments to artificially simulate various fault states of the de-icing system, acquiring abundant multi-sensor data under these fault conditions. This data provides a solid foundation for building a high-quality dataset, enabling the trained condition assessment and fault prediction models to more realistically reflect the actual system situation and possess stronger generalization ability and prediction accuracy.

[0046] 5. Optimize performance by integrating control strategies. Fully consider the zoned rotation control strategy of the de-icing system, incorporating relevant parameters (such as rotation cycle and zone working time) into data processing and model building. This not only helps to accurately assess the system's status under actual operating conditions but also enables targeted fault prediction and analysis based on the characteristics of the control strategy, thereby optimizing the de-icing system's performance, effectively reducing de-icing energy consumption, and improving the overall system operating efficiency.

[0047] 6. Enhanced Reliability Through Multi-Sensor Collaboration. Multiple sensors are deployed at various critical points in the de-icing system, including temperature and current sensors for the electric heating system, and acceleration and current sensors for the vibration system. This multi-sensor collaboration collects system operating data from different perspectives, providing rich information and avoiding misjudgments caused by single sensor failure or missing information. This significantly improves the reliability of condition assessment and fault prediction. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. The elements or parts in the drawings are not necessarily drawn to scale. Obviously, the drawings described below are some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0049] Figure 1 This is a schematic diagram of the structure of Embodiment 1 of the present invention;

[0050] Figure 2 This is a schematic diagram showing the layout of the de-icing zones;

[0051] Figure 3 This is a schematic diagram of the left wing after the de-icing section has been deployed.

[0052] Figure 4 This is a schematic diagram of the cycle composition in Embodiment 2 of the present invention;

[0053] Figure 5 This is a schematic diagram illustrating the composition of the partitioned working time in Embodiment 2 of the present invention;

[0054] Figure 6 This is a flowchart of Embodiment 2 of the present invention;

[0055] Figure 7 This is a flowchart of Embodiment 3 of the present invention;

[0056] Figure 8 The result of feature extraction for four round cycles of the change point detection algorithm;

[0057] Figure 9 This is a schematic diagram of residual overflow ice and de-icing areas (including icing areas and overflow ice areas) on the wings of a drone.

[0058] Reference numerals: 1 Skin, 2 Electric heating film, 3 Vibration exciter, 4 Structural component. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0060] In this document, suffixes such as "module," "part," or "unit" used to denote elements are used only for the purpose of illustrative purposes and have no specific meaning in themselves. Therefore, "module," "part," or "unit" may be used interchangeably.

[0061] In this document, the terms "upper," "lower," "inner," "outer," "front," "rear," "one end," and "the other end," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the present invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0062] In this document, unless otherwise explicitly specified and limited, the terms "installed," "equipped with," "connected," etc., should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, a direct connection, or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0063] In this document, "and / or" includes any and all combinations of one or more of the listed related items.

[0064] In this article, "multiple" means two or more, that is, it includes two, three, four, five, etc.

[0065] In this specification, certain embodiments may be disclosed in a range-bound format. It should be understood that this "range-bound" description is merely for convenience and brevity and should not be construed as a rigid limitation on the disclosed range. Therefore, the description of a range should be considered as having specifically disclosed all possible subranges and the individual numerical values ​​within those ranges. For example, a description of the range 1-6 should be considered as having specifically disclosed subranges such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6, etc., and the individual numbers within those ranges, such as 1, 2, 3, 4, 5, and 6. This rule applies regardless of the breadth of the range.

[0066] Example 1: As Figure 1 As shown, this embodiment provides a low-energy thermo-coupled de-icing system for aircraft with limited onboard energy, aiming to solve the aircraft de-icing problem under limited onboard energy conditions. By dividing the de-icing area into multiple zones and adopting a zone-based cyclic control method, combining heating and vibration de-icing methods, the system meets de-icing requirements while minimizing energy consumption. Furthermore, the system can automatically initiate de-icing operations based on environmental conditions, achieving intelligent de-icing. Its specific structure includes:

[0067] Multiple de-icing zones are symmetrically arranged along the fuselage axis; each de-icing zone is equipped with an independent heating device and a vibration device; the heating device is used to melt the ice layer and reduce the adhesion between the ice layer and the skin, and the vibration device is used to generate pulse vibrations on the skin to cause the ice layer on the skin to fall off.

[0068] The windward surfaces of the wings and tail are the first points where the aircraft comes into contact with airflow and moisture during flight. Icing often occurs first here, and the icing speed is faster and the ice layer accumulates thicker. Therefore, in this embodiment, the de-icing area is determined to be the windward surface of the wings and tail. The de-icing zone is the division of the de-icing area. Figure 1 The middle skin is located on the windward side of the wing or tail, which is where the de-icing section is located.

[0069] Specifically, the unfolded shape of the de-icing zones is rectangular (or parallelogram), and multiple de-icing zones are arranged on the windward side of the wing or tail and along the extension direction of the wing or tail. Since the wing and tail have upper and lower surfaces, the de-icing zones are set according to the cross-sectional shape of the wing or tail when applied to them. When the de-icing zones are unfolded and laid flat, they form a rectangle (or parallelogram), which facilitates the arrangement of multiple de-icing zones.

[0070] In addition, in this embodiment, the de-icing zones are arranged symmetrically, such as... Figure 2As shown, the left wing has n de-icing zones starting from L01 and L02, while the right wing has n de-icing zones starting from R01 and R02. The number and size of the de-icing zones are identical on both sides. Similarly, the de-icing zones T01 to T06 on the tail are also symmetrically arranged. This symmetrical arrangement of de-icing zones ensures symmetrical de-icing on both sides of the aircraft, thus avoiding any impact on the aircraft's aerodynamic performance.

[0071] More specifically, in this embodiment, the heating device uses an electric heating film laid on the inner side of the skin, and the area of ​​the electric heating film is consistent with the de-icing zone, achieving precise coverage of the heating area and the de-icing zone. Each de-icing zone can receive independent and targeted heating, avoiding the problem of excessive power demand due to an excessively large heating area, and also preventing the problem of poor de-icing effect due to insufficient heating.

[0072] In an electric heating film, the heating layer is composed of materials with good electrical conductivity and heat generation properties, such as metal heating wires and carbon nanotubes. When an electric current passes through the heating layer, electrical energy is converted into heat energy, causing the electric heating film to heat up rapidly. The heat is transferred to the ice layer through the skin, raising the temperature between the ice layer and the skin, thereby reducing the adhesion between the ice layer and the skin.

[0073] The insulating layer serves as an electrical isolation layer, separating the heating layer from the skin and other potentially contactable conductive components to prevent current leakage and safety issues. The insulating layer is typically made of high-temperature resistant materials with excellent insulation properties, such as polyimide and mica. The presence of the insulating layer ensures the electrical safety of the aircraft, preventing damage to the aircraft structure and equipment due to leakage from the electric heating film.

[0074] The protective layer is located on the outermost layer of the electric heating film, and its main function is to protect the heating layer and the insulation layer from external environmental corrosion and mechanical damage. The protective layer is usually made of materials with good wear resistance, corrosion resistance and flexibility, such as silicone rubber, fluoroplastics, etc.

[0075] In this embodiment, the vibration device includes a vibration exciter formed by winding a metal film, which is disposed between the skin and a structural component, such as a support frame inside the wing. Preferably, the vibration exciter is located at the center of each de-icing zone. When a pulsed current is applied, the metal film vibrates under the action of electromagnetic force. This design makes the vibration exciter compact and space-saving, suitable for installation within the limited space of an aircraft. Simultaneously, the winding method of the metal film can effectively control the direction and frequency of vibration, making the generated vibration more conducive to the removal of ice from the skin.

[0076] The vibration exciter is positioned between the skin and the structural component, using the structural component as a support. This allows the vibration to be transmitted more directly to the skin, reducing energy loss during the vibration transmission process.

[0077] The partition control module is used to perform partitioned rotation control of multiple de-icing partitions, so that multiple de-icing partitions take turns to perform de-icing work, and the de-icing partitions performing de-icing simultaneously are symmetrical along the fuselage axis.

[0078] Drones typically have limited onboard energy. During flight, a significant portion of their energy is consumed in critical flight operations, such as driving the propellers to generate thrust and controlling flight attitude. This results in relatively less power being allocated to the de-icing system. Using too much energy for de-icing would severely impact the drone's endurance and flight performance, and could even prevent the drone from completing its mission or returning safely due to insufficient power.

[0079] As a crucial heating component in de-icing systems, the electric heating film works by melting the ice layer on the skin using electric current (high temperature, continuous heating). This method requires significant energy consumption, typically necessitating high continuous power to achieve the desired de-icing effect. This patent uses an electric heating film that doesn't melt the entire ice layer. When de-icing is needed, electric heating reduces the adhesion between the ice layer and the skin (lower temperature, intermittent heating), followed by vibration to detach the ice layer. Further reducing power requirements, considering that each de-icing zone is equipped with an electric heating film of matching size, the power required for all de-icing zones to operate simultaneously would be extremely high. This remains a significant challenge for drone power supply systems with limited energy resources.

[0080] In this embodiment, in order to reduce the requirements of the UAV power supply system, a zoned de-icing strategy is adopted, that is, multiple de-icing zones are de-iced in turn in a certain order, which not only ensures the de-icing effect, but also reduces the requirements of the power supply system.

[0081] Each de-icing zone is equipped with an independent heating and vibration device. When it is the zone's turn to de-ic, the electric heating film first heats up to reduce the adhesion between the ice layer and the skin, and then the vibration exciter is activated to remove the ice layer. By reasonably setting the zone's working time and cycle, it can be ensured that each zone is fully de-iced, effectively removing the ice layer from the windward surfaces of the wings and tail.

[0082] The zoned de-icing strategy distributes the concentrated high power demand of existing technologies across different time periods by having multiple de-icing zones de-ic in turn. This significantly reduces instantaneous power demand, prevents power supply system failures due to overload, and ensures stable operation of the power supply system.

[0083] Furthermore, during the de-icing cycle, the de-icing zones operating simultaneously are symmetrical along the fuselage axis. In this embodiment, the de-icing zones are symmetrically distributed to ensure symmetrical de-icing during the process. For example, at the first moment, de-icing zone L01 on the left wing and de-icing zone R01 on the right wing begin de-icing simultaneously. At the second moment, de-icing zone L02 on the left wing and de-icing zone R02 on the right wing begin de-icing simultaneously, and so on. Multiple de-icing zones can also de-ic at the same time, as long as the principle of symmetry is followed. Of course, to minimize the demands on the power supply system, it is optimal to start two de-icing zones simultaneously.

[0084] The startup module is used to start the partition control module to control the de-icing partition to carry out de-icing work according to the preset de-icing conditions; the preset conditions are at least one of the following: the current ambient temperature is lower than the threshold temperature, the icing sensor detects icing, or the area is within an icing cloud.

[0085] In this embodiment, the de-icing system can be set to start automatically. The activation conditions include: 1. When the ambient temperature of the aircraft is below a set threshold temperature, such as -5°C, it indicates a risk of icing, and the system automatically initiates the de-icing procedure. 2. Icing sensors can monitor the aircraft surface for icing in real time. Once icing is detected, the system immediately starts the de-icing process. 3. When the aircraft is within an icing cloud, the likelihood of icing is extremely high, and the system automatically initiates the de-icing procedure to prepare for de-icing in advance and prevent ice accumulation.

[0086] In other embodiments, the system also includes sensors (e.g., the icing detector in CN110606209B) disposed in each de-icing zone for detecting the degree of icing (e.g., icing thickness), and accordingly, the zone control module is also used to dynamically adjust the queuing for each queuing cycle.

[0087] Specifically, before the end of the previous cycle, the partition control module obtains the degree of icing detected by the sensors on each de-icing partition and determines whether the degree of icing in each de-icing partition is greater than or equal to the preset icing threshold. If so, the de-icing partition is included in the current cycle queue (in the next cycle queue); otherwise, the de-icing partition is not included in the current cycle queue.

[0088] Furthermore, although the current icing level of some de-icing zones may not have reached the preset icing threshold, their waiting time (i.e., the time from the completion of their most recent de-icing to the current moment) is relatively long. Therefore, in the current cycle, their icing level is very likely to reach the preset icing threshold. Thus, in some other embodiments, if the icing level of a de-icing zone has not reached the preset icing threshold, but its waiting time has reached the preset time threshold, the de-icing zone will also be included in the current cycle queue; otherwise, the de-icing zone will not be included in the current cycle queue.

[0089] In other embodiments, the above-mentioned startup module is further configured to determine whether at least one de-icing partition adjacent to the current de-icing partition is in the current round-robin queue. If at least one de-icing partition adjacent to the current de-icing partition is in the current round-robin queue, the module determines whether the adjacent de-icing partition is the next de-icing partition to be de-iced. If the adjacent de-icing partition is the next de-icing partition to be de-iced, the module controls the vibration device of the adjacent de-icing partition to perform micro-vibration. If the adjacent de-icing partition has already completed de-icing, the module controls the heating unit of the adjacent de-icing partition to perform preheating.

[0090] While sensors can be used to monitor the degree of icing, the zoned, cyclical de-icing method employed in this application increases costs by requiring sensors to monitor the icing level in each de-icing zone. Furthermore, such sensors are typically mounted on the wing skin surface, which may not be desirable for any additional components in some scenarios. Therefore, in other embodiments, instead of using sensors on the icing zone skin surface for monitoring and dynamic adjustment, the method utilizes temperature sensors on the drone to monitor the ambient temperature. Different heating times are then set based on the ambient temperature, and the heating time is dynamically adjusted using real-time temperature data from each de-icing zone.

[0091] Although no sensors are installed to monitor the degree of icing, the ambient temperature of the drone varies, and consequently, the degree of icing on the wing skin may differ. Using the same heating time might not be sufficient to detach the interface layer between the skin and the ice layer in some low-temperature environments. Therefore, it is necessary to pre-calculate the heating time required to heat the skin to a preset temperature threshold at different ambient temperatures, achieving the detachment of the ice layer from the skin surface (not melting the entire ice layer). (For example, at -15°C, heating for 3 seconds reaches 30°C; while at -20°C, heating for 5 seconds reaches 35°C). Accordingly, temperature sensors are installed on the inner side of the skin in each de-icing zone to detect the temperature data after heating, and the heating time of the de-icing zone is dynamically adjusted based on the monitored temperature data. Specifically, the current ambient temperature is first obtained, and a preset heating device working time is matched in the database based on the current ambient temperature (as mentioned above, this can be pre-calculated through simulation, which is existing technology and will not be elaborated here). Then, the current temperature data of the current de-icing zone is obtained, and it is determined whether the temperature data of the de-icing zone is greater than or equal to the preset temperature threshold after the preset heating device working time. If so, the heating device is controlled to stop heating; otherwise, the heating device is controlled to continue heating until the preset temperature threshold is reached. The actual heating working time of the heating device is recorded, and this actual heating working time is used as the heating device working time of the de-icing zone under the current ambient temperature.

[0092] Furthermore, since the actual heating time is longer than the preset initial heating device working time, the aforementioned start-up module is also used to determine whether the difference Δt between the actual heating time and the preset initial heating device working time is greater than or equal to the aforementioned vibration delay time t. delay If it is less than the vibration delay time t delay Then shorten the vibration delay time t delay The heating time of the heating device is used as part of the vibration extension time; if it is greater than or equal to the vibration extension time, the sum of the actual heating time and the vibration device working time is used as the partition working time of the de-icing zone. Accordingly, the partition working time interval between the current de-icing zone and the next de-icing zone is shortened while the cycle period remains unchanged.

[0093] In other embodiments, the above-mentioned startup module is further configured to identify at least one de-icing zone adjacent to the current de-icing zone when heating the current de-icing zone, and control the vibration device of the adjacent de-icing zone that has completed de-icing to vibrate; at the same time, control the heating unit of the adjacent de-icing zone that will be the next to be de-iced to be preheated.

[0094] Example 2:

[0095] This embodiment also provides a control method for a low-energy de-icing system with limited airborne energy, applied to the aforementioned de-icing system, see [link to relevant documentation]. Figure 6 The control method specifically includes:

[0096] S1 defines the de-icing zones according to the maximum de-icing power provided by the power supply system.

[0097] In this embodiment, given a fixed area for the de-icing section, the number of de-icing zones is primarily determined by the maximum de-icing power provided by the power supply system. Too few de-icing zones result in a high power requirement per zone, increasing the load on the power supply system. Conversely, too many de-icing zones increase control complexity and reduce de-icing efficiency.

[0098] Therefore, theoretically, the maximum de-icing power that the power supply system can provide should be used as the upper limit of the power for simultaneously operating de-icing zones, i.e., Q ≥ q * s * n, where Q is the maximum de-icing power, q is the power per unit area of ​​the electric heating film, and n is the number of simultaneously operating de-icing zones. Generally, n=2 is preferable for the number of simultaneously operating zones. While ensuring symmetrical de-icing, the number of de-icing zones should be minimized. Additionally, it is worth noting that q needs to ensure that the skin reaches the specified temperature (e.g., 40℃) within a preset time.

[0099] Therefore, we can deduce the maximum area of ​​a single de-icing zone, i.e., s ≤ Q / (q*n). After obtaining the area of ​​a single de-icing zone, the number of de-icing zones can be obtained from the total area of ​​the de-icing area.

[0100] Other examples Figure 3 As shown, the unfolded shape of the de-icing zone in this embodiment is rectangular, with parameters including length L and width M. The length L is a preset value, and the width M is set according to the maximum area s of the de-icing zone.

[0101] More specifically, the length of the de-icing zone is calculated using the formula L=L0*K, where L is the length of the de-icing zone, L0 is the ideal length of the de-icing zone, and K is the redundancy coefficient. The ideal length L0 of the de-icing zone can be obtained through physical simulation, model calculation, etc. The redundancy coefficient K ranges from 1.2 to 1.4. Since the ice in the de-icing zone is pre-melted using a heating film before de-icing with a vibration device, some of the liquid formed during the ice melting process may flow with the airflow to the rear of the icing area on the wing, forming overflow ice. The overflow ice has different effects on the wing's aerodynamic performance depending on its location on the wing. For example, overflow ice near the rear of the icing area has a greater impact on the wing's aerodynamic performance. Therefore, the impact of different locations on the wing's aerodynamic performance is obtained in advance through physical simulation, wind tunnel implementation, and model calculation, thus obtaining the redundancy coefficient K. This allows the overflow ice area and the area prone to ice formation to be divided into a single de-icing zone. See [link to relevant documentation]. Figure 3 and Figure 9 This allows the de-icing process to remove not only the ice buildup on the leading edge of the wing, but also some of the overflow ice formed after the ice melts, and even prevents the formation of overflow ice in that area, thus avoiding the impact of overflow ice formed after the ice melts on the aerodynamic performance of the wing.

[0102] Once the maximum area s of the de-icing zone and the length L of the de-icing zone are determined, the width M of the de-icing zone can be determined.

[0103] S2 sets a cycle, each cycle including several partition working times; the number of partition working times is at least 1 / 2 of the number of de-icing partitions, so that at least two de-icing partitions symmetrically arranged along the fuselage axis work simultaneously during each partition working time; there is an interval between the partition working times.

[0104] The core of the low-energy de-icing system in this embodiment lies in the adoption of a zoned round-robin control strategy, which can effectively reduce de-icing energy consumption. A schematic diagram of the zoned round-robin control strategy is shown below. Figure 4 As shown, a high level indicates that the partition is working, and a low level indicates that the partition is not working. Taking the five de-icing system partitions on the left wing of the aircraft as an example, after L01 works for a period of time (the time length is the partition working time ts), partition L02 starts working, and so on. The time from the start of L01 working to the start of the next L01 working is the cycle period Tc, and the time from the start of the current partition to the start of the next partition is the partition working time interval Δts. All parameters can be set according to the actual situation.

[0105] It is foreseeable that, since the de-icing work in this embodiment needs to be carried out symmetrically, while the left wing de-icing zone L01 is performing de-icing work, the right wing de-icing zone R01 is also performing de-icing work. And when the left wing de-icing zone L02 is performing de-icing work, the right wing de-icing zone R02 is also performing de-icing work, and so on.

[0106] In some embodiments, to save energy, the cycle queue for the next cycle is dynamically adjusted based on the current icing level of each icing zone before the start of each cycle or near the end of the previous cycle. Specifically, a corresponding sensor is set up in each de-icing zone to detect the icing level of each zone, and then it is determined whether the icing level of each de-icing zone is greater than or equal to a preset icing threshold (i.e., the current icing is very thin or sparse, and de-icing is not required in this cycle). If so, the de-icing zone is added to the cycle de-icing queue; otherwise, it is not included in the cycle de-icing queue. Figure 4 For example, if the data detected by the sensors indicates that the icing level of de-icing zone L04 is less than the preset icing threshold, then de-icing zone L04 will be removed from the round-robin queue until its icing level is detected to be greater than or equal to the preset icing threshold, at which point it will be added back to the round-robin queue. Accordingly, during this round-robin de-icing process, once de-icing zone L03 is completed, de-icing of de-icing zone L05 will proceed directly.

[0107] Furthermore, although the current icing level of some de-icing zones may not have reached the preset icing threshold, their waiting time (i.e., the time from the completion of their most recent de-icing to the current moment) is relatively long. Therefore, in the current cycle, their icing level is very likely to reach the preset icing threshold. Thus, in some other embodiments, if the icing level of a de-icing zone has not reached the preset icing threshold, but its waiting time has reached the preset time threshold, the de-icing zone will also be included in the current cycle queue; otherwise, the de-icing zone will not be included in the current cycle queue.

[0108] However, installing the aforementioned sensors in each de-icing zone increases costs, and in some scenarios, such as smaller drones, it is generally undesirable to install any additional components on the skin surface of the drone's wings. Therefore, without installing sensors on the de-icing zones to monitor the degree of icing in each zone, this embodiment also provides another control method. Specifically, by installing a temperature sensor on the inner side of the skin of each de-icing zone to detect the temperature data of the skin after heating, the working time of the heating device in that de-icing zone is dynamically adjusted based on the monitored temperature data. Specifically, the current ambient temperature is first obtained, and a preset heating device working time is matched in the database based on the current ambient temperature (as mentioned above, this can be pre-calculated through simulation, which is existing technology and will not be elaborated here). Then, the current temperature data of the current de-icing zone is obtained, and it is determined whether the temperature data of the de-icing zone after the preset heating device working time is greater than or equal to a preset temperature threshold. If so, the heating device is controlled to stop heating; otherwise, the heating device is controlled to continue heating until the preset temperature threshold is reached. The actual heating working time of the heating device is recorded, and this actual heating working time is used as the heating device working time of the de-icing zone at the current ambient temperature.

[0109] Furthermore, since the actual heating time is longer than the preset initial heating device working time, the aforementioned start-up module is also used to determine whether the difference Δt between the actual heating time and the preset initial heating device working time is greater than or equal to the aforementioned vibration delay time t. delay If it is less than the vibration delay time t delay Then shorten the vibration delay time t delay The heating time of the heating device is used as part of the vibration extension time; if it is greater than or equal to the vibration extension time, the sum of the actual heating time and the vibration device working time is used as the partition working time of the de-icing zone. Accordingly, the partition working time interval between the current de-icing zone and the next de-icing zone is shortened while the cycle period remains unchanged.

[0110] In addition, the interval Δts between zones can be set according to the icing conditions. When the ambient temperature is low and the icing rate is fast, the interval Δts can be shortened or even reduced to 0, thereby improving de-icing efficiency. When the icing rate is low, the interval Δts can be extended, thereby reducing energy consumption.

[0111] The time required for each de-icing zone to operate is related to the number of de-icing zones and the number of de-icing zones allowed to operate simultaneously. Since this embodiment requires symmetrical de-icing of the wings or tail, at least two zones must operate simultaneously, and the corresponding time required for each zone should be at least half the number of de-icing zones. For example, if four zones are allowed to operate simultaneously, then the time required for each zone is one-quarter of the number of zones. That is, if m zones are allowed to operate simultaneously (where m is an even number), then the time required for each zone to operate is 1 / m of the number of zones.

[0112] Of course, in other embodiments, even if the cycle is dynamically adjusted based on the degree of icing, the working time of the heating device in the corresponding de-icing zone can be further dynamically adjusted based on the monitored temperature data in each cycle.

[0113] In addition, in this embodiment, the de-icing zones on the wings are de-iced sequentially from the inside out; the de-icing layers on the tail are de-iced sequentially from the outside in.

[0114] In some embodiments, the partition working time includes the heating device working time, the delay time, and the vibration device working time, which are sequentially arranged.

[0115] A schematic diagram of the single de-icing zone control strategy is shown below. Figure 5 As shown, the working time of a single zone is equal to the working time t of the electric heating system. h Vibration system operating time t v and vibration delay time t delay The sum of the values ​​during the heating device's operating time t h Inside, a direct current of a certain current and voltage is applied to the electric heating film, causing the surface temperature of the wing skin to rise, with a delay time t. delay This is the time from the moment the heating device is disconnected to the moment the vibration device starts operating; it can be set according to actual conditions and can be set to 0. Subsequently, the vibration system starts working, and the electrostatic excitation coil generates vibration force under the action of pulsed DC current, causing ice on the skin surface to fall off. Figure 4 The diagram shows a pulse signal of five vibrations.

[0116] Set the delay time t delay The reason is that the electric heating film is laid on the inside of the skin, and the heat needs to be conducted to the ice layer on the outer surface through the skin material. The skin itself has a certain thickness and thermal resistance, and it takes time for heat to be transferred from the inside to the outside ice layer. If the vibration device is started immediately after heating, the temperature at the interface between the ice layer and the skin may not have reached the critical value, resulting in the ice layer not being sufficiently loosened and the vibration de-icing effect being poor. Delay time t delayThis provides a buffer time for heat conduction, ensuring that the heat from the heating layer is evenly diffused to the outer surface of the skin, creating a sufficient temperature gradient at the interface between the ice layer and the skin, effectively reducing adhesion. Preferably, this t delay The duration is 30 seconds to 1 minute.

[0117] Furthermore, as mentioned earlier, since overflow ice is formed when the ice on the leading edge of the wing is heated and melted, and the melted liquid flows to the rear under the action of airflow, there is a certain time lag in the formation of overflow ice. If the vibration device is activated immediately after heating, it may change the direction of liquid flow and cause unpredictable effects (e.g., reverse flow or diffusion at the leading edge of the wing). On the other hand, if the fluid has not flowed to the rear or overflow ice has not yet formed, activating the vibration device at this time would waste energy. In other words, by setting this delay time, on the one hand, it provides a buffer time for heat conduction, ensuring that the heat from the heated layer is evenly diffused to the outer surface of the skin, creating a sufficient temperature gradient at the interface between the ice layer and the skin, effectively reducing adhesion; on the other hand, it provides a buffer time for the flow of liquid formed by the melting of ice, allowing the melted liquid to diffuse to the rear or near the rear of the de-icing area, or when most of the liquid is far from the center of the de-icing area, before activation. This not only accelerates the flow of liquid to the residual overflow ice area to a certain extent, avoiding the probability of overflow ice forming in the de-icing area, but also saves energy to a certain extent.

[0118] The de-icing effect of the low-energy de-icing system is mainly related to the skin surface temperature and vibration magnitude. The skin surface temperature can be controlled by the electric heating working time and electric heating current, while the vibration magnitude can be adjusted by the pulse voltage and pulse width. In this embodiment, the selection principles of the core parameters of the low-energy de-icing system control strategy are shown in Table 1:

[0119] Table 1 Selection Strategy for Core Parameters

[0120]

[0121] S3 uses a cyclic control system to de-ice the heating and vibration devices.

[0122] In some embodiments, the low-energy de-icing system can achieve self-starting of the de-icing system in the following ways:

[0123] 1. Ambient temperature < set ambient temperature, such as -5℃;

[0124] 2. The icing sensor detected icing conditions.

[0125] 3. Based on feedback from other airborne image equipment algorithms, the aircraft is currently in an icing cloud layer.

[0126] In existing technologies, separate vibration units are used to remove overflow ice from specific areas. However, for drones with limited onboard energy, this increases both the drone's weight and system energy consumption. Compared to large aircraft, small drones with limited onboard energy are prone to overflow ice formation after the ice melts at the wing leading edge. However, due to the structural characteristics of such aircraft wings, not all overflow ice will negatively impact the wing's aerodynamics. In this application, when dividing the de-icing zones, a redundancy coefficient K (preferably 1.2-1.4) is set based on the wing's structural characteristics and experimental experience. The area of ​​each de-icing zone is calculated based on this redundancy coefficient K, and the heating film area is set accordingly. This allows the heating film to melt the ice in each de-icing zone, preventing the melted ice from overflowing in areas far from the wing leading edge and affecting the wing's aerodynamics. Furthermore, the vibration units are used to ensure timely de-icing. In other words, this application only targets the spilled ice that needs to be removed, in addition to removing accumulated ice (i.e., the spilled ice near the wing leading edge that affects the wing's aerodynamics, see [reference]). Figure 9 De-icing is performed, while unnecessary overflow ice (i.e., overflow ice that has little impact on wing aerodynamics, see [reference]) is removed. Figure 9 (Residual overflow ice) does not require unnecessary operations, greatly reducing energy consumption.

[0127] Because a rotating mechanism is used to de-ice each de-icing zone, and there is a time interval between two adjacent de-icing operations, when de-icing the current de-icing zone, water formed during the de-icing process may flow to adjacent de-icing zones due to airflow caused by aircraft turning or other reasons. However, since the adjacent de-icing zone may have already completed de-icing, or even if it is the next de-icing zone, there is still a time interval before de-icing it, ice may reform between adjacent adjacent de-icing zones in low-temperature environments, resulting in incomplete de-icing or increasing the difficulty of de-icing the next de-icing zone. Therefore, to reduce the probability of this happening, based on the same inventive concept, this invention also provides another control method, which includes the steps of the above embodiments, except that when the current de-icing zone is heated, it also includes the step of:

[0128] S301A identifies at least one de-icing zone adjacent to the current de-icing zone. If the adjacent de-icing zone has already completed de-icing, proceed to step S302A. If the adjacent de-icing zone is the next de-icing zone to be de-iced, proceed to step S303A.

[0129] S302A controls the vibration devices of adjacent de-icing zones to vibrate.

[0130] S303A controls the heating units of adjacent de-icing zones to preheat.

[0131] As mentioned earlier, since all de-icing zones on the same wing are rotated sequentially, the current de-icing zone may be adjacent to one or two other de-icing zones. Furthermore, due to the different locations of the current de-icing zone, its adjacent de-icing zones may fall into two categories: those that are about to be de-iced and those that have already been de-iced. Therefore, different measures need to be adopted to reduce the probability of ice forming between two adjacent de-icing zones.

[0132] by Figure 4 For example, if the current de-icing zone is L03, and its two adjacent de-icing zones L02 and L04 are both in the current round-robin queue, and de-icing zone L02 has already completed its de-icing work, but de-icing zone L04 is the next zone to be de-iced. Therefore, when de-icing zone L03 is undergoing heating and de-icing, the water it produces may flow to de-icing zones L02 and / or L04. Since de-icing zone L04 itself requires heating before de-icing and already has a certain thickness of ice, it can be preheated simultaneously during the de-icing process of de-icing zone L03 (i.e., the heating unit corresponding to de-icing zone L04 is turned on for heating; preferably, the heating power during preheating is less than the heating power during de-icing). As for the de-icing zone L02, since it has already undergone de-icing (i.e., its surface has no ice layer or a very thin ice layer), its corresponding vibration device can be used to apply slight vibration, thereby changing the direction of the water flow, for example, directing it towards the overflow area, or making the water flow in this zone thinner. Preferably, the vibration power at this time is less than the vibration power during the de-icing process.

[0133] In other embodiments, since the round-robin queue is dynamically adjusted in advance based on the degree of icing, the process of de-icing in the current de-icing zone also includes the following steps:

[0134] S301B determines whether at least one de-icing partition adjacent to the current de-icing partition is in the current round-robin queue. If yes, proceed to step S302B; otherwise, proceed to step S303B.

[0135] As mentioned earlier, since the circulation queue is adjusted in advance based on the degree of icing in each de-icing zone, the adjacent de-icing zones of the current de-icing zone may not be in the current circulation queue, or they may be in the current circulation queue. Therefore, different measures need to be taken for these two situations to reduce the probability of ice forming between two adjacent de-icing zones.

[0136] S302B determines whether the adjacent de-icing zone is the next de-icing zone to be de-iced. If so, proceed to step S304; otherwise, proceed to step S303.

[0137] S303 controls the vibration devices in adjacent de-icing zones to vibrate.

[0138] S304 controls the heating units of adjacent de-icing zones to preheat.

[0139] Example 3: Figure 7 As shown, this embodiment provides a method for real-time status assessment and fault prediction of a de-icing system, applied to the thermo-coupled de-icing system in Embodiment 1. The specific steps include:

[0140] S101 simulates the normal operation, open circuit, and system failure states of the thermally coupled de-icing system through ground condition monitoring tests to obtain sensor data under each state.

[0141] The operating data of the electric heating device in the thermocoupled de-icing system includes temperature and current.

[0142] More specifically, a temperature sensor and a current sensor for monitoring the current of the electric heating device are installed on the inside of the skin of each de-icing zone. The current sensor is a Hall effect current sensor, which means that the wire passes through the measuring hole of the corresponding sensor and has no effect on the electric heating circuit.

[0143] The operating data of the vibration device in the thermally coupled de-icing system include vibration frequency and current.

[0144] More specifically, an acceleration sensor is installed on the skin of each de-icing zone near the vibration exciter or near the exciter support, and a current sensor can also be installed to monitor the vibration pulse current.

[0145] After installing each sensor, ground condition monitoring tests (e.g., conducting ground condition monitoring tests on the prototype on the leading edge section of a wing with 5 de-icing zones) can be conducted to artificially simulate the fault conditions of the de-icing system and obtain multi-sensor data corresponding to different fault conditions.

[0146] Specifically, in the ground condition monitoring experiment, the principle of the control method in the above embodiment 2 is adopted, and a cycle control mechanism is used to obtain sensor data in each cycle, and the data is preprocessed and used as the dataset for training the model.

[0147] For example, when the electric heating and / or vibration devices in each de-icing zone are in an open-circuit state, the sensor data in each cycle includes, for example, the temperature data monitored by the temperature sensor, the current data monitored by the current sensor, and the acceleration data monitored by the accelerometer. Because the power supply is automatically cut off very quickly once the circuit is open, the corresponding sensor data cannot be collected, thus appearing as an open circuit in the data representation.

[0148] For example, under normal operating conditions, the sensor data of the electric heating device and / or vibration device in each de-icing zone during each cycle, such as temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the acceleration sensor.

[0149] For example, under fault conditions, the sensor data for each cycle of the electric heating device and / or vibration device in each de-icing zone includes: temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the accelerometer. Fault conditions of the electric heating device include: partial open circuit, power attenuation, response delay, and thermocouple failure. Fault conditions of the vibration device include: abnormal vibration acceleration magnitude.

[0150] Of course, in addition to ground condition monitoring experiments, simulation experiments can also be used to obtain the dataset needed for training.

[0151] In some embodiments, the sheer volume of data collected by sensors makes training extremely difficult. For example, with a 120-second cycle, a temperature sampling rate of 50Hz, and current and acceleration sampling rates of 5000Hz, five cycles would generate approximately 500MB of data. In one cycle, the data length for a single electric heating channel would be 120 * 5000 = 600,000 data points; for five zone current sensors, this would amount to 10 data points. The same applies to vibration. This makes training the model very challenging.

[0152] Therefore, in this embodiment, the raw data collected by the sensor is not used for training. Instead, it is preprocessed to extract the time-domain and frequency-domain features in each cycle. A dataset is then constructed based on the extracted time-domain and frequency-domain features, and training is performed on it.

[0153] Preferably, the time-domain features include: the working time (i.e., electric heating time) of each electric heating device within a cycle under normal working conditions, open circuit conditions, or fault conditions; the mean, variance, and standard deviation of the electric heating current of each heating device within its working time (i.e., calculated based on the current data monitored by the current sensor, the mean, variance, and standard deviation of the current of each heating device within the corresponding working time); and the maximum, minimum, and mean temperatures monitored by the sensors on all de-icing zones within a cycle (i.e., selecting the de-icing zones with the highest and lowest temperatures based on the temperature data monitored by the temperature sensor, and calculating the mean temperature of all de-icing zones that are de-icing in this cycle), etc.

[0154] Frequency domain characteristics include: the natural frequency, excitation frequency, vibration frequency, and amplitude of the vibration device (i.e., the natural frequency, excitation frequency, vibration frequency, and amplitude of each vibration device in each cycle are calculated based on the acceleration data monitored by the accelerometer).

[0155] The above preprocessing significantly reduces the amount of data to be processed.

[0156] S102 uses the dataset constructed in step S101 to train the state assessment model and the fault prediction model.

[0157] In some embodiments, the training state evaluation model and fault prediction model can be trained using existing machine learning models, such as Gaussian Mixture Model (GMM) and Support Vector Data Description (SVDD), to obtain the state evaluation model.

[0158] Preferably, the status assessment involves preprocessing the dataset collected when a de-icing cycle is completed and then inputting it into the status assessment module to determine the current working status.

[0159] Preferably, the fault prediction model is prepared by preprocessing historical data from multiple completed cycle periods and then inputting it into a trained fault prediction model for fault prediction. For example, in the case of a poorly soldered joint with excessive resistance, during operation, the continuous flow of current at the solder joint causes it to melt due to the high resistance and excessive heat generation. The resistance continues to increase, eventually leading to an open circuit. This will be reflected in the data, such as current and temperature data. While we may not fully understand the cause of some abnormal states at present, anomalies in the data can indicate that there is definitely an anomaly, either a sensor malfunction or a problem with a component of the anti-icing and de-icing system.

[0160] In some embodiments, when constructing the dataset, to simplify the data, a 2×n state matrix is ​​constructed as label samples of the system states. Each label sample corresponds to a set of time-domain features and frequency-domain features, which are extracted after preprocessing the sensor data collected in one cycle under the corresponding state. Here, n is the number of partitions corresponding to the electric heating device or vibration device. In the row representing the state of the electric heating device, each element represents a partition corresponding to the electric heating device. The value of the element 0 represents the electric heating device being open-circuited, the value of the element 1 represents the electric heating device being normal, and the value of the element 2 represents the electric heating device being abnormal (i.e., fault, for example, aging of the electric heating film). In the row representing the state of the vibration device, each element represents a vibration partition. The value of the element 0 represents the vibration device being open-circuited, the value of the element 1 represents the vibration system being normal, and the value of the element 2 represents the vibration device being abnormal (i.e., the vibration device is in a fault state).

[0161] For example, the state matrix ;

[0162] The first row represents the status of the electric heating device, and the second row represents the status of the vibration device, each containing 5 partitions (i.e., 5 columns per row). Therefore, the status matrix means that the electric heating device in partition 3 has an open circuit; the vibration device in partition 1 is malfunctioning; the vibration device in partition 4 has an open circuit; and the vibration devices in the remaining partitions are functioning normally.

[0163] The state assessment and fault prediction model described in this invention is trained on a large dataset. The labels of the dataset are state matrices, and the corresponding data are multi-sensor data or features obtained under the corresponding states. By mapping the state of the electric heating device or vibration device to a state matrix, the complexity of data processing is simplified.

[0164] The dataset described in this invention requires data preprocessing, such as data cleaning, feature extraction, and data normalization. Data cleaning involves handling outliers and missing values ​​in the sensor data; feature extraction needs to be combined with the actual de-icing system's control strategy, such as cycle time extraction, electric heating time, and pulse count.

[0165] Furthermore, the state assessment model in this invention takes into account real-time multi-sensor data, including time-series signals such as temperature, current, and vibration acceleration, and outputs a 2×n state matrix of the current de-icing system. This matrix is ​​used to express, for example, whether a certain section of the electric heating system is open-circuited or whether a certain section of the vibration system is abnormal. Applicable models include Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN / LSTM), Fully Connected Neural Networks (FCN), and Graph Neural Networks (GNN).

[0166] Preferably, the state assessment model evaluates the probability of various failures of the electric heating device / vibration device in the current partition based on the input multi-sensor data, takes the one with the highest probability as the assessment result, and then constructs the state matrix based on the assessment result.

[0167] The fault prediction model in this invention takes historical sensor data sequences, such as temperature and current fluctuations over the past 10 minutes, as input. It outputs a fault probability and / or fault state matrix for a future period and provides maintenance suggestions, such as "Data in partition 3 shows a trend of xx; please check and maintain partition 3." Applicable models include LSTM / Transformer, Gradient Boosting Tree (GBM), Isolation Forest, and survival analysis models.

[0168] S103 installs sensors on aircraft equipped with thermally coupled de-icing systems to collect operational data from the de-icing system.

[0169] By installing sensors on the aircraft that are identical to those used in ground tests, consistency between the real-time collected data and the training dataset is ensured, providing reliable input for online model inference. The types and installation methods of the sensors are described in step S101 and will not be repeated here.

[0170] Since this invention targets a zoned de-icing system, it is necessary to identify the cycle period. Therefore, this embodiment also provides a method for identifying the cycle period, the specific steps of which include:

[0171] S1031 collects temperature sensor data for a single zone of the thermally coupled de-icing system over multiple cycle periods.

[0172] S1032 filters out the starting point of the electric heating system in each cycle from the temperature sensor data to obtain the variable point data point sequence.

[0173] When an electric heating system starts up, it causes a significant temperature change, typically manifested as a rapid temperature rise. The starting point of this temperature change is called the operating start point, or change point. These change points are identified from temperature sensor data using a specific change point detection algorithm (existing technology, not detailed here). This results in a change point data point sequence. The change point data point sequence records the moment of each start-up of the electric heating system. These moments are key nodes for calculating the cycle time. Each change point corresponds to one start-up of the electric heating system, and its data point index value contains time information.

[0174] Taking the temperature sensor data (sampling rate 10Hz) of a single section of a thermocoupled de-icing system during four cycles of electric heating operation as an example, a change-point detection algorithm is used to identify the cycle, i.e., the time interval between the first operation and the next operation of that section. In the ground test, the cycle was 30 seconds, and the change-point detection results are as follows: Figure 8 As shown, the specific variable point data results are: [175, 480, 785, 1085, 1395]. Dividing the data points by the sampling rate gives the time, so the detection cycle for the four rounds is [30.5, 30.5, 30, 31], which is very close to the true value of 30s.

[0175] S1033 calculates the polling cycle based on the time interval between adjacent variable data points.

[0176] The polling cycle refers to the time interval between the start of one operation and the start of the next operation for a given partition. The time interval between adjacent variable data points precisely reflects this cycle. Since there is a fixed sampling frequency when collecting data, the actual time interval, i.e., the polling cycle, can be calculated based on the difference in index values ​​of the variable data points and the sampling frequency.

[0177] Specifically, using the formula:

[0178] ;

[0179] Calculate the cycle period; where T i For the i-th cycle, D i Let D be the index value of the i-th working starting point. i+1 Here is the index value of the (i+1)th working starting point, and m is the sampling frequency. For example, see... Figure 8 If the sampling frequency is 10Hz and the index difference between two adjacent variable data points is 300, then the polling period is 300 / 10=30 seconds.

[0180] S104 inputs the data returned by the sensor into the trained state assessment model and fault prediction model to assess the real-time state of the de-icing system and predict possible faults.

[0181] The data transmitted from the sensors comes from various sensors deployed on the aircraft's de-icing system, including temperature and current sensors monitoring the electric heating system, and acceleration and current sensors monitoring the vibration system. The data is real-time and continuously reflects the operating status of the de-icing system. Through a predetermined transmission channel, this data is fed into pre-trained condition assessment and fault prediction models.

[0182] The state assessment model determines the real-time status of the de-icing system based on input sensor data. During the training phase, the model learns the characteristics of sensor data under various fault states and their corresponding state matrices. Upon receiving real-time data, the model compares and analyzes the features of the current data with those in the training set to identify the current state of the de-icing system. For example, it can determine whether a section of the electric heating system is open-circuited, or whether the vibration system is functioning normally, and outputs the assessment results in the form of a state matrix, allowing operators to intuitively understand the real-time operating status of the de-icing system.

[0183] The fault prediction model focuses on utilizing historically collected multi-sensor data to uncover potential patterns and trends, enabling proactive assessments of potential faults in the de-icing system. This model considers that changes in certain parameters during long-term operation may foreshadow impending failures. For example, by monitoring current trends and temperature rise rates in the electric heating system, or abnormal fluctuations in the vibration frequency of the vibration system, combined with pre-fault data patterns learned during training, the model predicts the types of faults that may occur in the system in the near future, the probability of occurrence, and the areas that may be affected. It then provides corresponding maintenance recommendations to help ground personnel prepare for maintenance in advance and reduce flight safety risks caused by de-icing system failures.

[0184] In other embodiments, based on the above-described inventive concept, state assessment or fault prediction can be performed separately. Specifically, if state assessment is performed separately, it includes steps S101-104. The difference is that in step S102, only the state assessment model is trained. Correspondingly, in step S104, the data collected by the sensor during the completion of one de-icing cycle is preprocessed and then input into the trained state assessment model to obtain the real-time state assessment result of the de-icing system. Similarly, if fault prediction is performed separately, it includes steps S101-S104. The difference is that in step S102, only the fault prediction model is trained. Correspondingly, in step S104, the historical data of all cycles returned by the sensor during the multiple completed cycles is preprocessed and then input into the trained fault prediction model to obtain the fault prediction result of the de-icing system.

[0185] Example 4: Based on the method of Example 3 above, the present invention also provides a real-time status assessment and fault prediction system for a de-icing system, comprising:

[0186] The fault status data acquisition module is used to simulate the fault status of the de-icing system through ground condition monitoring tests of the thermally coupled de-icing system and acquire sensor data under different fault statuses.

[0187] The model training module is used to build a dataset using sensor data under different fault conditions to train the state assessment model and the fault prediction model.

[0188] The working data acquisition module is used to install sensors on aircraft equipped with de-icing systems to collect working data of the de-icing system;

[0189] The evaluation and prediction module is used to input the data returned by the sensors into the trained state evaluation model and fault prediction model to evaluate the real-time state of the de-icing system and predict possible faults.

[0190] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0191] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a computer terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0192] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for real-time status assessment and fault prediction of a de-icing system, characterized in that, The de-icing system includes: multiple de-icing zones symmetrically arranged along the fuselage axis; each de-icing zone is equipped with an independent heating device and a vibration device; the heating device is used to melt the ice layer and reduce the adhesion between the ice layer and the skin, and the vibration device is used to generate pulse vibrations on the skin to cause the ice layer on the skin to fall off; a zone control module is used to set the cycle period and perform zone cycle control on the multiple de-icing zones, so that the multiple de-icing zones take turns to perform de-icing work, and the de-icing zones performing de-icing simultaneously are symmetrical along the fuselage axis; wherein, the multiple de-icing zones are defined according to the maximum de-icing power provided by the power supply system; each cycle includes several zone working times; the number of zone working times is at least 1 / 2 of the number of de-icing zones, so that at least two de-icing zones symmetrically arranged along the fuselage axis work simultaneously during each zone working time; there is an interval between the zone working times; the zone working time includes the working time of the heating device, a delay time, and the working time of the vibration device set sequentially; accordingly, The real-time status assessment and fault prediction method for the de-icing system includes: A ground condition monitoring test of the de-icing system simulates the normal operation, open circuit, and fault conditions of the de-icing system, acquiring sensor data under each condition. The sensor data includes: temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the accelerometer under normal operation of the electric heating device and vibration device in each cycle; temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the accelerometer under fault conditions of the electric heating device and vibration device in each cycle; and temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the accelerometer under open circuit conditions of the electric heating device / vibration device in each cycle. Sensor data under normal, open-circuit, and fault states are preprocessed to obtain a training dataset. This dataset is then used to train a state assessment model and a fault prediction model. Specifically, the preprocessing includes extracting the time-domain and frequency-domain features of the sensor data collected in each cycle, and constructing a 2×n state matrix as state labels. Each state matrix corresponds to a set of time-domain and frequency-domain features. Here, n represents the number of partitions in the electric heating device or vibration device. In the row representing the state of the electric heating device, each element represents a partition for electric heating, and the value of the element represents the state of the electric heating device. In the row representing the state of the vibration device, each element represents a partition for vibration, and the value of the element represents the state of the vibration device. Sensors are installed on aircraft equipped with de-icing systems to collect operational data from the de-icing systems. The data transmitted back from the sensors is input into the trained state assessment model and fault prediction model to assess the real-time state of the de-icing system and predict possible faults.

2. The method for real-time status assessment and fault prediction of a de-icing system according to claim 1, characterized in that: The operating data of the electric heating device in the de-icing system includes temperature and current; the operating data of the vibration device in the de-icing system includes vibration frequency, vibration amplitude, and current.

3. The method for real-time status assessment and fault prediction of a de-icing system according to claim 2, characterized in that: For the electric heating device in the de-icing system, the temperature of the skin is monitored by a temperature sensor installed inside the skin, and the current is monitored by a current sensor. For the vibration device in the de-icing system, the vibration frequency and vibration amplitude are monitored by an acceleration sensor installed on the skin near the vibration exciter or at the exciter support, and the current is monitored by a current sensor.

4. The method for real-time status assessment and fault prediction of a de-icing system according to claim 1, characterized in that... Multiple de-icing zones are located on the windward side of the wings or tail and arranged along the direction of extension of the wings or tail.

5. The method for real-time status assessment and fault prediction of a de-icing system according to claim 1, characterized in that... The steps for defining de-icing zones based on the maximum de-icing power provided by the power supply system include: The maximum area of ​​a single de-icing zone is calculated using the formula s≤Q / (q*n), where s is the maximum area of ​​the de-icing zone, Q is the maximum de-icing power, q is the power of the electric heating film per unit area, and n is the number of de-icing zones working simultaneously; where q satisfies the requirement that the skin reaches the specified temperature within a preset time.

6. The method for real-time status assessment and fault prediction of a de-icing system according to claim 1, characterized in that: In the 2×n state matrix; In the row representing the status of the electric heating device, the value of an element is 0, which means the electric heating device is open-circuited; the value of an element means the electric heating system is normal; and the value of an element means the electric heating device is abnormal. In the row representing the status of the vibration device, an element with a value of 0 indicates that the vibration device is open-circuited, an element with a value of 1 indicates that the vibration system is normal, and an element with a value of 2 indicates that the vibration device is abnormal.

7. The method for real-time status assessment and fault prediction of a de-icing system according to claim 1, characterized in that... The steps for identifying the cycle of the de-icing system include: Collect temperature sensor data for a single zone in the de-icing system during multiple cycle periods; The starting point of the electric heating device in each cycle is selected from the temperature sensor data to obtain the variable point data point sequence; The polling cycle is calculated based on the time interval between adjacent variable data points.

8. The method for real-time status assessment and fault prediction of a de-icing system according to claim 7, characterized in that... Using the formula: ； Calculate the cycle period; where T i For the i-th cycle, D i Let D be the index value of the i-th working starting point. i+1 is the index value of the (i+1)th working starting point, and m is the sampling frequency.

9. The method for real-time status assessment and fault prediction of a de-icing system according to claim 1, characterized in that: Before training the state assessment model and the fault prediction model, the dataset is preprocessed; the preprocessing includes data cleaning, feature extraction and data normalization.

10. A real-time status assessment and fault prediction system for a de-icing system, characterized in that... include: The fault status data acquisition module is used to simulate the fault status of the de-icing system through ground status monitoring tests of the de-icing system and acquire sensor data under different fault statuses. The sensor data includes: temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the accelerometer sensor under normal operating conditions of the electric heating device and the vibration device in each cycle; temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the accelerometer sensor under fault conditions of the electric heating device and the vibration device in each cycle; and temperature data monitored by the temperature sensor, current data monitored by the current sensor, and acceleration data monitored by the accelerometer sensor under open-circuit conditions of the electric heating device / vibration device in each cycle. The model training module is used to preprocess sensor data under normal, open-circuit, and fault states to obtain a training dataset, and then use this dataset to train a state assessment model and a fault prediction model. Specifically, the preprocessing includes extracting the time-domain and frequency-domain features of the sensor data collected in each cycle, and constructing a 2×n state matrix as state labels, with each state matrix corresponding to a set of time-domain and frequency-domain features; where n is the number of partitions in the electric heating system or vibration system; in the row representing the state of the electric heating system, each element represents an electric heating partition, and the value of the element represents the state of the electric heating device; in the row representing the state of the vibration system, each element represents a vibration partition, and the value of the element represents the state of the vibration device. The working data acquisition module is used to install sensors on aircraft equipped with de-icing systems to collect working data of the de-icing system; The evaluation and prediction module is used to input the data returned by the sensors into the trained state evaluation model and fault prediction model to evaluate the real-time state of the de-icing system and predict possible faults. The de-icing system includes: Multiple de-icing zones are symmetrically arranged along the fuselage axis. Each de-icing zone is equipped with an independent heating device and a vibration device. The heating device is used to melt the ice layer and reduce the adhesion between the ice layer and the skin. The vibration device is used to generate pulse vibrations on the skin to cause the ice layer on the skin to fall off. The multiple de-icing zones are determined in advance according to the maximum de-icing power provided by the power supply system. The partition control module is used to set the rotation cycle and perform partition rotation control on multiple de-icing partitions, so that the multiple de-icing partitions take turns to perform de-icing work, and the de-icing partitions performing de-icing simultaneously are symmetrical along the fuselage axis; wherein, each rotation cycle includes a number of partition working times; the number of partition working times is at least 1 / 2 of the number of de-icing partitions, so that at least two de-icing partitions arranged symmetrically along the fuselage axis work simultaneously during each partition working time; there are intervals between the partition working times; The working time of the partition includes the working time of the heating device, the delay time, and the working time of the vibration device, which are set sequentially.

Images