Engine pre-ignition control methods, vehicle-side, cloud-side, media and products
By detecting the risk of engine pre-ignition at the vehicle end and uploading the data to the cloud to generate personalized protection strategies, the problem of insufficient accuracy in pre-ignition protection in existing technologies is solved, and precise protection and resource optimization for engine pre-ignition are achieved.
Patent Information
- Application Number
- CN202611526978.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-29
- Publication Date
- 2026-11-17
AI Technical Summary
In existing technologies, engine pre-ignition protection strategies cannot accurately adapt to the actual operating environment and fuel quality of different vehicles, resulting in insufficient protection accuracy, and the high-frequency data upload increases resource consumption.
By continuously collecting engine operating data on the vehicle, the risk of pre-ignition is detected and uploaded to the cloud when a risk is detected. The cloud generates a pre-ignition risk map and identification model, and provides personalized engine protection strategies. It also uses deep learning models to mine pre-ignition patterns in massive amounts of data and updates the strategy when the engine is turned off and the vehicle is stopped.
It improves the precision of engine pre-ignition protection, reduces data transmission volume and resource consumption, and adapts to changes in different regions and fuel quality, thus extending engine life.
Smart Images

Figure CN123109891A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of engine control technology, specifically to a method for controlling engine pre-ignition, vehicle-side components, cloud-side components, media, and products. Background Technology
[0002] Pre-ignition is one of the key bottlenecks restricting engine power performance and durability. Especially in turbocharged direct injection engines, which often operate under high load and low speed conditions, this region is a high-risk area for low-speed pre-ignition. Pre-ignition is initiated by high-temperature deposits or incandescent particles, resulting in a high pressure rise rate, which can potentially damage critical components such as pistons, connecting rods, and valves. Therefore, it is necessary to promptly identify and effectively prevent the risk of pre-ignition in such engines.
[0003] In related technologies, engine operating data from multiple vehicles are simply summarized and trend statistically analyzed. A universal engine optimization strategy is generated based on the global average operating status and distributed in batches to all vehicles to implement pre-ignition protection.
[0004] However, using only superficial data statistics cannot uncover the complex patterns of pre-ignition hidden in massive vehicle data that cannot be summarized by human experience, and the strategy output is a one-size-fits-all approach, resulting in insufficient accuracy in pre-ignition protection. Summary of the Invention
[0005] In view of this, this application aims to provide a method, device, medium and product for controlling engine pre-ignition, so as to solve the problem of insufficient accuracy of pre-ignition protection in traditional related technologies.
[0006] The first aspect of this application provides a method for controlling engine pre-ignition, applied to a vehicle, comprising: continuously collecting engine operating data; detecting the presence of pre-ignition risk based on the engine operating data; when a pre-ignition risk is detected, sending vehicle information and engine operating data corresponding to the pre-ignition risk to the cloud, so that the cloud generates a pre-ignition risk map based on vehicle information from multiple vehicles and engine operating data corresponding to the pre-ignition risk; obtaining a trained pre-ignition recognition model based on engine operating data corresponding to the pre-ignition risk from multiple vehicles; generating an engine pre-ignition protection strategy corresponding to the current vehicle based on the pre-ignition risk map and the pre-ignition recognition model; receiving the engine pre-ignition protection strategy sent from the cloud; and performing pre-ignition protection on the engine according to the engine pre-ignition protection strategy.
[0007] In one possible implementation of this application, the continuous collection of engine operating data includes: collecting engine operating data at a first sampling frequency when no risk of pre-ignition is detected; and switching to a second sampling frequency to collect engine operating data when a risk of pre-ignition is detected; wherein the second sampling frequency is higher than the first sampling frequency.
[0008] In one possible implementation of this application, sending engine operating data corresponding to the pre-ignition risk to the cloud includes: taking the moment when the pre-ignition risk is detected as a baseline, extracting engine operating data within a preset time window before the moment and engine operating data collected within a preset time window after the moment from the collected engine operating data; and sending the engine operating data within the two preset time windows to the cloud.
[0009] In one possible implementation of this application, sending engine operating data within two preset time windows to the cloud includes: compressing the engine operating data within the two preset time windows; and sending the compressed engine operating data to the cloud during network idle periods.
[0010] In one possible implementation of this application, the vehicle information includes at least a vehicle identification number (VIN) and a vehicle location; sending the vehicle information to the cloud includes: performing irreversible hashing on the VIN in the vehicle information; converting the vehicle location in the vehicle information into a regional location; and removing personal information from the vehicle information; desensitizing the processed vehicle information; and sending the desensitized vehicle information to the cloud; wherein, the regional location is used by the cloud to collect engine operation data from multiple vehicles and generate a pre-ignition risk map; and the VIN is used for engine pre-ignition protection strategies generated for different vehicles.
[0011] In one possible implementation of this application, receiving an engine pre-ignition protection strategy sent from the cloud includes: receiving the current engine pre-ignition protection strategy from the vehicle via an over-the-air download channel when the vehicle is in a shutdown state; wherein the engine pre-ignition protection strategy carries a digital signature generated by the cloud for the engine pre-ignition protection strategy; verifying the digital signature of the engine pre-ignition protection strategy; and storing the engine pre-ignition protection strategy received this time after the digital signature verification is successful.
[0012] In one possible implementation of this application, after performing pre-ignition protection on the engine, the method further includes: returning to the step of continuously collecting engine operating data, continuously collecting engine operating data and detecting pre-ignition risk, and when a pre-ignition risk is detected, sending vehicle information and engine operating data corresponding to the pre-ignition risk to the cloud, so that the cloud iteratively updates the pre-ignition risk map, the pre-ignition identification model and the engine pre-ignition protection strategy.
[0013] A second aspect of this application also provides a method for controlling engine pre-ignition, applied in the cloud, comprising: receiving vehicle information and engine operating data sent by a vehicle terminal; wherein the vehicle terminal continuously collects engine operating data; detecting pre-ignition risk based on the engine operating data; when a pre-ignition risk is detected, sending the vehicle information and the engine operating data corresponding to the pre-ignition risk to the cloud; generating a pre-ignition risk map based on the vehicle information and the engine operating data corresponding to the pre-ignition risk from multiple vehicle terminals; obtaining a trained pre-ignition recognition model based on the engine operating data corresponding to the pre-ignition risk from multiple vehicle terminals; generating an engine pre-ignition protection strategy corresponding to the current vehicle terminal based on the pre-ignition risk map and the pre-ignition recognition model; and sending the engine pre-ignition protection strategy to the vehicle terminal so that the vehicle terminal can perform pre-ignition protection on the engine according to the engine pre-ignition protection strategy.
[0014] In one possible implementation of this application, the vehicle information includes at least a regional location; the engine operating data corresponding to the pre-ignition risk includes at least the time when the pre-ignition risk occurs; and a pre-ignition risk map is generated based on the vehicle information from multiple vehicle terminals and the engine operating data corresponding to the pre-ignition risk, including: based on the regional location of each vehicle terminal, the time when the pre-ignition risk occurs, and the fuel quality corresponding to the regional location, the probability of pre-ignition occurring under different regions, different times, and different fuel qualities is calculated to generate a pre-ignition risk map.
[0015] In one possible implementation of this application, a pre-ignition identification model is obtained based on engine operating data corresponding to pre-ignition risks from multiple vehicle terminals. This includes: using a federated learning framework to collaboratively train the pre-ignition identification model using engine operating data from multiple vehicle terminals; during the collaborative training process, mining pre-ignition precursor features in the engine operating data to obtain a global pre-ignition identification model, and using the global pre-ignition identification model as the pre-trained pre-ignition identification model.
[0016] In one possible implementation of this application, a pre-ignition protection strategy for the engine corresponding to the current vehicle is generated based on a pre-ignition risk map and a pre-ignition identification model. This includes: obtaining the engine load and engine speed of the current vehicle; combining the current vehicle's regional location, the time of pre-ignition risk occurrence, and the fuel quality corresponding to the regional location to obtain the pre-ignition probability from the pre-ignition risk map; correcting the ignition advance angle corresponding to the engine load and engine speed in the baseline ignition advance angle upper limit pulse spectrum table based on the pre-ignition probability to obtain the ignition advance angle upper limit pulse spectrum table for the current vehicle; determining the adjustment range of the engine control parameters for the current vehicle based on the pre-ignition probability; combining the ignition advance angle upper limit pulse spectrum table and the adjustment range of the engine control parameters into engine control parameters; and using the model parameters of the pre-ignition identification model and the engine control parameters as the engine pre-ignition protection strategy for the current vehicle.
[0017] In one possible implementation of this application, before sending the engine pre-ignition protection strategy to the vehicle, the method further includes: inputting the engine pre-ignition protection strategy into the engine pre-ignition digital twin model, so that the engine pre-ignition digital twin model performs simulation verification of the engine pre-ignition protection strategy to obtain the pre-ignition risk prediction result corresponding to the engine pre-ignition protection strategy; and adjusting the engine pre-ignition protection strategy according to the pre-ignition risk prediction result.
[0018] A third aspect of this application provides a control device for engine pre-ignition, applied to a vehicle, comprising: a data acquisition module for continuously acquiring engine operating data; a detection module for detecting the presence of pre-ignition risk based on the engine operating data; a first transmission module for transmitting vehicle information and engine operating data corresponding to the pre-ignition risk to the cloud when a pre-ignition risk is detected, so that the cloud generates a pre-ignition risk map based on vehicle information from multiple vehicles and engine operating data corresponding to the pre-ignition risk; and obtains a trained pre-ignition recognition model based on the engine operating data corresponding to the pre-ignition risk from multiple vehicles; and generates an engine pre-ignition protection strategy corresponding to the current vehicle based on the pre-ignition risk map and the pre-ignition recognition model; a first receiving module for receiving the engine pre-ignition protection strategy transmitted from the cloud; and a pre-ignition protection module for performing pre-ignition protection on the engine according to the engine pre-ignition protection strategy.
[0019] A fourth aspect of this application also provides a control device for engine pre-ignition, applied in the cloud, comprising: a second receiving module for receiving vehicle information and engine operating data sent by a vehicle terminal; wherein the vehicle terminal continuously collects engine operating data; detects pre-ignition risk based on the engine operating data; and when a pre-ignition risk is detected, sends the vehicle information and the engine operating data corresponding to the pre-ignition risk to the cloud; a first generating module for generating a pre-ignition risk map based on vehicle information from multiple vehicle terminals and the engine operating data corresponding to the pre-ignition risk; an acquisition module for acquiring a trained pre-ignition recognition model based on the engine operating data corresponding to the pre-ignition risk from multiple vehicle terminals; a second generating module for generating an engine pre-ignition protection strategy corresponding to the current vehicle terminal based on the pre-ignition risk map and the pre-ignition recognition model; and a second sending module for sending the engine pre-ignition protection strategy to the vehicle terminal so that the vehicle terminal can perform pre-ignition protection on the engine according to the engine pre-ignition protection strategy.
[0020] The fifth aspect of this application provides a vehicle end, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to cause the at least one processor to perform an engine pre-ignition control method as described in the first aspect and possible implementations of the first aspect.
[0021] The sixth aspect of this application provides a cloud platform, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to cause the at least one processor to perform an engine pre-ignition control method as described in the second aspect and possible implementations thereof.
[0022] The seventh aspect of this application provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement an engine pre-ignition control method as described in the first aspect and its possible implementations, as well as the second aspect and its possible implementations.
[0023] The eighth aspect of this application provides a computer program product comprising: a computer program that, when executed by a processor, implements an engine pre-ignition control method as described in the first aspect and its possible implementations, and the second aspect and its possible implementations.
[0024] The method, device, medium, and product for controlling engine pre-ignition provided in this application involve continuously collecting engine operating data; detecting the presence of pre-ignition risk based on the engine operating data; when a pre-ignition risk is detected, sending vehicle information and the corresponding engine operating data to the cloud; enabling the cloud to generate a pre-ignition risk map based on vehicle information from multiple vehicle terminals and the corresponding engine operating data; obtaining a trained pre-ignition recognition model based on the engine operating data corresponding to the pre-ignition risks from multiple vehicle terminals; generating an engine pre-ignition protection strategy for the current vehicle terminal based on the pre-ignition risk map and the pre-ignition recognition model; receiving the engine pre-ignition protection strategy sent from the cloud; and performing pre-ignition protection on the engine according to the engine pre-ignition protection strategy. By continuously collecting engine operating data and detecting pre-ignition risk at the vehicle terminal, and only uploading vehicle information and the corresponding engine operating data when a risk is detected, and having the cloud generate a pre-ignition risk map and a trained pre-ignition recognition model based on multi-vehicle data, it is possible to uncover complex pre-ignition patterns in massive amounts of vehicle data and generate personalized engine protection strategies for the current vehicle terminal, thereby improving the accuracy of engine pre-ignition protection. Attached Figure Description
[0025] To more clearly illustrate the technical solutions in the specific embodiments or related technologies of this application, the drawings used in the description of the specific embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0026] Figure 1 A flowchart illustrating the engine pre-ignition control method provided in this application embodiment. Figure 1 .
[0027] Figure 2 A flowchart illustrating the engine pre-ignition control method provided in this application embodiment. Figure 2 .
[0028] Figure 3 A flowchart illustrating the engine pre-ignition control method provided in this application embodiment. Figure 3 .
[0029] Figure 4 A schematic diagram of the engine pre-ignition control device provided in the embodiments of this application. Figure 1 .
[0030] Figure 5 A schematic diagram of the engine pre-ignition control device provided in the embodiments of this application. Figure 2 .
[0031] Figure 6 This is a schematic diagram of the vehicle-side hardware structure provided in an embodiment of this application.
[0032] Figure 7 This is a schematic diagram of the cloud hardware structure provided in an embodiment of this application. Detailed Implementation
[0033] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0034] In the field of engine pre-ignition protection technology, to control the risk of engine pre-ignition, related technologies typically summarize and statistically analyze the engine operating data of multiple vehicles, generate a universal engine optimization strategy based on the global average operating status, and then distribute it in batches to all vehicles for pre-ignition protection. This approach attempts to guide vehicle-side pre-ignition protection by aggregating operating data from multiple vehicles and leveraging group statistical patterns.
[0035] However, when this solution is applied to operating scenarios under different regions, seasons, and fuel qualities, its performance is less than ideal. A fundamental contradiction lies in the fact that the application of related technologies to multi-vehicle operating data remains at a superficial statistical level, failing to uncover the complex patterns of pre-ignition hidden within massive amounts of vehicle data that cannot be summarized by human experience. Simultaneously, its strategy output is a one-size-fits-all approach, without differentiated adaptation to the actual operating environment, conditions, and fuel quality of different vehicles, resulting in insufficient accuracy in pre-ignition protection. Specifically, under high-load, low-speed conditions, the time required for pre-ignition to progress from abnormal combustion to pressure increase is short. If the uniform strategy fails to identify the risk in time, components such as pistons, connecting rods, or valves may already be under high pressure. Under high-temperature or low-octane fuel conditions, the uniform optimization strategy generated based on the global average state is mismatched with the current operating environment, easily leading to response delays or misjudgments. Requiring all vehicles to continuously upload high-frequency data to improve accuracy would increase the consumption of data transmission, storage, and communication resources.
[0036] Through in-depth analysis, the inventors discovered that the root cause of the aforementioned contradiction lies in the fact that while the relevant technologies aggregate data from multiple vehicles, they are only used for superficial summarization and trend statistics, failing to establish a mechanism for learning the complex patterns of pre-ignition from massive amounts of data and generating protection strategies tailored to individual vehicle differences. Because the global average operating status masks the differences between different vehicles, it is difficult to identify early signs of pre-ignition coupled with fuel quality, regional climate, and engine operating conditions. Furthermore, the batch distribution of uniform strategies cannot be dynamically adjusted according to the actual operating status of individual vehicles, resulting in vehicle-side pre-ignition protection capabilities remaining at a global experience level, making precise adaptation difficult.
[0037] To address the aforementioned technical problems, this application proposes the following technical concept: The vehicle continuously collects engine operating data, detects the presence of pre-ignition risk based on this data, and when a pre-ignition risk is detected, sends vehicle information and the corresponding engine operating data to the cloud. The cloud generates a pre-ignition risk map based on vehicle information from multiple vehicles and the corresponding engine operating data, and obtains a trained pre-ignition recognition model based on the engine operating data from multiple vehicles. Then, the cloud uses the pre-ignition risk map and the pre-ignition recognition model to generate a pre-ignition protection strategy for the current vehicle. The vehicle receives this pre-ignition protection strategy and performs pre-ignition protection on the engine. This embodiment solves the problems in related technologies where only shallow data statistics are used, failing to uncover complex pre-ignition patterns and resulting in insufficient accuracy in pre-ignition protection due to a lack of unified strategy output.
[0038] Figure 1 A flowchart illustrating the engine pre-ignition control method provided in this application embodiment. Figure 1 This embodiment describes an interaction scenario between the vehicle and the cloud, such as... Figure 1 As shown, the details are as follows: S101: The vehicle continuously collects engine operating data.
[0039] S102: The vehicle detects the risk of pre-ignition based on engine operating data.
[0040] S103: When the vehicle detects a risk of pre-ignition, it sends vehicle information and engine operating data corresponding to the pre-ignition risk to the cloud.
[0041] S104: The cloud generates a pre-ignition risk map based on vehicle information from multiple vehicle terminals and engine operating data corresponding to the pre-ignition risk.
[0042] S105: The cloud-based system obtains a trained pre-ignition identification model based on engine operating data corresponding to pre-ignition risks from multiple vehicle terminals.
[0043] S106: The cloud generates the engine pre-ignition protection strategy corresponding to the current vehicle based on the pre-ignition risk map and pre-ignition identification model.
[0044] S107: The cloud sends the engine pre-ignition protection strategy to the vehicle.
[0045] S108: The vehicle end performs pre-ignition protection for the engine according to the engine pre-ignition protection strategy.
[0046] In subsequent embodiments, the implementation process of each step on the vehicle side and the cloud side is described.
[0047] Figure 2 A flowchart illustrating the engine pre-ignition control method provided in this application embodiment. Figure 2 In this embodiment, the execution entity is the vehicle end. For example... Figure 2 As shown, the method includes: S201: Continuously collect engine operating data.
[0048] In this embodiment, engine operating data may include multimodal sensor data (such as cylinder pressure, vibration, temperature, and pressure sensor data), operating condition data (such as engine speed, load, air-fuel ratio, and ignition advance angle), and control commands (such as fuel injection quantity, ignition timing, and variable valve timing (VVT) control signals). The fusion of these three types of data enables a comprehensive reconstruction of the dynamic process before and after the occurrence of pre-ignition risk.
[0049] In this embodiment, the collected engine operation data is cached locally, which can cache data from the past 3 years.
[0050] In some embodiments, different data acquisition strategies may be employed. Specifically, when no risk of pre-ignition is detected, engine operating data is acquired at a first sampling frequency; when a risk of pre-ignition is detected, the acquisition of engine operating data is switched to a second sampling frequency; wherein the second sampling frequency is higher than the first sampling frequency.
[0051] The first sampling frequency can be 100–200Hz, covering the main frequency changes of engine operating parameters. It can capture the evolution trends of key parameters such as engine speed, load, and air-fuel ratio, maintaining basic monitoring of engine operating status. A lower sampling frequency can reduce local cache usage and data processing computation, enabling the on-board system to operate continuously for a long time without affecting the real-time performance of the vehicle's electrical / electronic architecture (EEA).
[0052] The second sampling frequency can be 5–10 kHz, covering the main frequency components of combustion pressure fluctuations, and can accurately record high-frequency characteristics such as cylinder pressure waveforms and vibration spectra before and after the occurrence of pre-ignition risk. High-frequency sampling is only initiated when a pre-ignition risk is detected, avoiding the storage and bandwidth pressure caused by long-term high-frequency sampling, and achieving on-demand acquisition. The second sampling frequency mentioned here is higher than the first sampling frequency.
[0053] In this embodiment, the dynamic switching of the sampling frequency achieves a balance between continuous monitoring and accurate capture. Under the constraint of limited communication bandwidth, it maintains continuous monitoring of engine operating data while accurately capturing key information such as high-frequency combustion waveforms when the risk of pre-ignition occurs. Compared to a fixed sampling frequency scheme, the data transmission volume is reduced by more than 90% while preserving the complete dynamic process.
[0054] S202: Based on engine operating data, detect whether there is a risk of pre-ignition.
[0055] In this embodiment, the likelihood of pre-ignition occurring, either currently or soon, is determined based on collected engine operating data. Detection methods may include: rule-based judgment based on multi-sensor thresholds, or inference using a locally deployed pre-ignition identification model on the engine operating data.
[0056] In this embodiment, the risk of pre-ignition can be detected through collaborative prediction using a local Artificial Intelligence (AI) model and multi-sensor thresholds. The local AI model can uncover pre-ignition precursor features in engine operating data, such as cylinder pressure fluctuation patterns and abnormal vibration spectra, while multi-sensor thresholds provide physical constraints. This allows the detection of pre-ignition risk to possess both data-driven pattern recognition capabilities and meet the constraints of physical mechanisms, thereby improving detection reliability.
[0057] S203: When a pre-ignition risk is detected, the vehicle information and the engine operating data corresponding to the pre-ignition risk are sent to the cloud so that the cloud can generate a pre-ignition risk map based on the vehicle information and engine operating data corresponding to the pre-ignition risk from multiple vehicle terminals; and obtain a trained pre-ignition recognition model based on the engine operating data corresponding to the pre-ignition risk from multiple vehicle terminals; and generate an engine pre-ignition protection strategy corresponding to the current vehicle terminal based on the pre-ignition risk map and the pre-ignition recognition model.
[0058] In this embodiment, when a risk of pre-ignition is detected, the vehicle information sent from the vehicle to the cloud may include the Vehicle Identification Number (VIN), vehicle location, vehicle type, and engine model.
[0059] In this embodiment, the engine operating data corresponding to the pre-ignition risk can include the moment when the pre-ignition risk occurs, as well as the engine operating data within a preset time window before and after that moment. Data before the occurrence of the pre-ignition risk records the evolution of pre-ignition precursor characteristics, while subsequent data records the system response and damage propagation process after the pre-ignition occurs. This complete time window data enables the cloud to analyze the pre-ignition triggering mechanism and influencing factors, providing causal chain information for the pre-ignition identification model and improving its ability to identify pre-ignition precursor characteristics.
[0060] In this embodiment, the pre-ignition risk map generated in the cloud can include the probability of pre-ignition occurrence in terms of geographical, temporal, and fuel dimensions. The geographical dimension reflects the impact of factors such as altitude and ambient temperature in different regions on pre-ignition; the temporal dimension reflects the modulating effect of seasonal changes and diurnal temperature variations on pre-ignition risk; and the fuel dimension reflects the differences in fuel quality (such as octane rating) at different gas stations. The three-dimensional risk map can dynamically adjust the engine pre-ignition protection strategy according to the current vehicle environment, significantly improving the adaptability of the engine pre-ignition protection strategy to environmental changes.
[0061] In this embodiment, the pre-trained pre-ignition detection model acquired from the cloud can be a deep learning model based on Long Short-Term Memory (LSTM) networks or Transformers. LSTM networks can capture temporal dependencies in engine operating data and uncover the dynamic evolution patterns of pre-ignition precursor features; Transformer models have stronger long-distance dependency modeling capabilities and can correlate complex interactions between multimodal sensor data. Compared to traditional threshold judgment methods, deep learning models can discover high-dimensional nonlinear patterns that are difficult to summarize manually, significantly improving the accuracy and robustness of pre-ignition risk prediction.
[0062] In this embodiment, the cloud-generated engine pre-ignition protection strategy can include model parameters of the pre-ignition detection model and engine control parameters. The model parameters of the pre-ignition detection model, such as neural network weights, enable the continuous evolution of the vehicle-side local AI model; the engine control parameters, such as the ignition advance angle upper limit pulse spectrum table and engine control parameter adjustment range, directly constrain the actions of the actuators, quickly triggering protective measures when a pre-ignition risk is detected. The collaborative updating of model parameters and control parameters achieves a closed loop of prediction and protection, improving pre-ignition detection capabilities while ensuring the executable nature of the protection strategy.
[0063] S204: Receive engine pre-ignition protection strategy sent from the cloud.
[0064] S205: Based on the engine pre-ignition protection strategy, the engine is protected against pre-ignition.
[0065] In this embodiment, engine pre-ignition protection strategies can include: limiting the ignition advance angle based on the ignition advance angle upper limit pulse spectrum table; adjusting the fuel injection quantity according to the adjustment range of engine control parameters; limiting engine torque output; increasing engine speed; and adjusting the ignition advance angle retraction. Specifically, retraction of the ignition advance angle reduces peak combustion pressure and pressure rise rate, directly suppressing pre-ignition; adjusting the fuel injection quantity changes the air-fuel mixture concentration, preventing the formation of incandescent particles in locally rich areas; limiting torque output and increasing engine speed remove the engine from the high-risk zone of low-speed pre-ignition. These multi-dimensional control measures work synergistically to minimize the risk of pre-ignition while ensuring power performance, thus extending engine lifespan.
[0066] In this embodiment, when receiving and deploying the engine pre-ignition protection strategy sent from the cloud, it is necessary to ensure the security, integrity and reliability of the strategy update process.
[0067] Specifically, when the vehicle is in a shutdown state, the engine pre-ignition protection strategy is received from the vehicle via an over-the-air download channel. The engine pre-ignition protection strategy carries a digital signature generated by the cloud for the engine pre-ignition protection strategy. The engine pre-ignition protection strategy is digitally signed and verified. After the digital signature verification is successful, the received engine pre-ignition protection strategy is stored.
[0068] In this embodiment, when the engine is off and parked, it stops running. Updating the engine pre-ignition protection strategy at this time will not interfere with the engine's current control logic, avoiding transient disturbances caused by strategy switching. The Over-The-Air (OTA) channel, typically a 4G / 5G cellular network, provides a direct connection to the cloud, eliminating the need for manual operation by the vehicle owner and enabling silent policy updates and continuous optimization throughout the entire lifecycle.
[0069] In this embodiment, the hash value of the engine pre-ignition protection strategy is encrypted using a private key in the cloud, and the digital signature is verified by the cloud's public key at the vehicle end. The digital signature ensures the trustworthiness of the strategy's origin, prevents man-in-the-middle attacks or malicious strategy injection, and safeguards vehicle functional safety.
[0070] In this embodiment, the vehicle executes a digital signature verification algorithm locally, comparing the received hash value with the hash value after decrypting the digital signature. The engine pre-ignition protection strategy is considered legitimate only if the two match.
[0071] In this embodiment, the engine protection policy that has passed digital signature verification replaces the old version of the engine protection policy. The vehicle loads the new engine protection policy upon the next startup, achieving persistent policy updates.
[0072] As described above, the process involves continuously collecting engine operating data; detecting the risk of pre-ignition based on this data; and when a pre-ignition risk is detected, sending vehicle information and the corresponding engine operating data to the cloud. The cloud then generates a pre-ignition risk map based on vehicle information from multiple vehicle terminals and the corresponding engine operating data. It also acquires a trained pre-ignition recognition model based on this data. Finally, it generates a pre-ignition protection strategy for the current vehicle based on the risk map and the recognition model. The system receives the pre-ignition protection strategy from the cloud and implements pre-ignition protection for the engine. By continuously collecting engine operating data and detecting pre-ignition risks at the vehicle terminal, and only uploading vehicle information and corresponding engine operating data when a risk is detected, and having the cloud generate a pre-ignition risk map and a trained pre-ignition recognition model based on multi-vehicle data, the complex patterns of pre-ignition in massive amounts of vehicle data can be uncovered. This allows for the generation of personalized engine protection strategies for the current vehicle, improving the accuracy of pre-ignition protection.
[0073] Furthermore, in some embodiments, a time window-based strategy can be used when the vehicle uploads data to the cloud. Specifically, based on the moment when a pre-ignition risk is detected, engine operating data within a preset time window before that moment and engine operating data within a preset time window after that moment are extracted from the collected engine operating data; the engine operating data within the two preset time windows are then sent to the cloud.
[0074] The data prior to the occurrence of pre-ignition risk records the engine's evolution from a normal combustion state to a pre-ignition risk state, including the dynamic trajectory of pre-ignition precursor characteristics. This data enables the cloud-based pre-ignition identification model to learn "what combination of operating conditions and what evolution path will lead to pre-ignition," significantly improving the model's ability to identify pre-ignition precursors.
[0075] The data following the occurrence of pre-ignition risk records the evolution of the combustion process, the response of the control system (such as ignition angle retraction and fuel injection adjustment), and the status of the mechanical system (such as vibration amplitude and temperature changes). This data enables the cloud to assess the actual effectiveness of the current protection strategy, the severity of the pre-ignition event, and the potential damage, providing feedback for strategy optimization.
[0076] In this embodiment, symmetrical time window data centered on the moment when the risk of premature combustion occurs fully records the entire process from the precursor of premature combustion to the occurrence of premature combustion to the system response. This enables the cloud to establish a causal relationship between the triggering factors of premature combustion and the effect of the system response, providing high-quality labeled samples for the training of the premature combustion identification model and the optimization of protection strategies.
[0077] For example, both the preset time window before and after this moment can be 5 seconds. A 5-second time window corresponds to approximately 167 combustion cycles of the engine under typical operating conditions (such as 2000 rpm), which is sufficient to cover the evolution of pre-ignition precursor characteristics (which typically accumulates over dozens of cycles) and the transient response process after pre-ignition occurs.
[0078] As described above, based on the moment when a pre-ignition risk is detected, engine operating data within a preset time window before and after the detected moment are extracted from the collected engine operating data. This data is then sent to the cloud. By extracting and uploading operating data from two preset time windows centered on the moment the pre-ignition risk is detected, the cloud can obtain complete data before and after the occurrence of the pre-ignition risk, which helps to accurately identify pre-ignition characteristics and train the pre-ignition recognition model.
[0079] In order to achieve efficient transmission of large-capacity, high-frequency data under the constraint of limited communication bandwidth, in some embodiments, the engine operation data within two preset time windows is compressed; and the compressed engine operation data is sent to the cloud during network idle periods.
[0080] Engine operating data, especially high-frequency sensor data such as cylinder pressure and vibration, has strong time-series correlation and redundancy. Using a dedicated time-series data compression algorithm, a compression ratio of 8:1 to 15:1 can be achieved. This enables the onboard communication module (Telematics BOX, T-Box) to complete data uploads within the limited bandwidth of 4G / 5G cellular networks, avoiding storage overflow caused by data backlog.
[0081] For example, compression can employ a time-series data compression algorithm based on LZ4. The LZ4 compression algorithm boasts extremely high compression / decompression speeds and moderate compression ratios, typically achieving 8:1 to 12:1 for time-series data, making it suitable for the real-time processing needs of automotive embedded systems. Compared to general-purpose compression algorithms, LZ4 has lower CPU and memory requirements. Compared to dedicated wavelet compression, LZ4's lossless nature ensures the integrity of data received in the cloud, meeting the quality requirements for subsequent model training.
[0082] For example, network idle periods can be when the vehicle is turned off and parked, or during overnight charging. When the vehicle is turned off and parked, the vehicle's electronic system enters a low-power mode, and the T-Box can operate independently to complete data uploads without affecting other vehicle systems; during overnight charging, the vehicle is stationary and powered by an external power source, so data uploads do not consume battery power.
[0083] In this embodiment, the above-mentioned data compression and transmission strategy enables the vehicle to achieve efficient transmission of large-capacity, high-frequency data under the constraint of limited communication bandwidth. Compared with the original data upload, the data transmission volume is reduced by 90%, the transmission latency is reduced by 80%, and network resources are avoided during the driver's normal driving hours.
[0084] Building upon the above embodiments, in this embodiment, considering that vehicle-cloud collaboration involves the collection and transmission of vehicle data, privacy protection is a crucial prerequisite for the technology's implementation. The vehicle information includes at least the vehicle identification number (VIN) and the vehicle's location.
[0085] In this embodiment, when sending vehicle information to the cloud, the vehicle identification number in the vehicle information is subjected to irreversible hashing; the vehicle location in the vehicle information is converted into a regional location; and personal information in the vehicle information is removed; the processed vehicle information is desensitized, and the desensitized vehicle information is sent to the cloud.
[0086] Among them, the regional location is used to collect engine operation data from multiple vehicles in the cloud and generate a pre-ignition risk map; the vehicle identification number is used to generate engine pre-ignition protection strategies for different vehicles.
[0087] In this embodiment, the irreversible hash algorithm can be SHA-256. The vehicle identification number (VIN) is mapped to a fixed-length hash value, such as a 64-bit hexadecimal string. This mapping process is unidirectional and cannot be reversed to derive the original VIN. The cloud only stores the hash value, which can identify multiple data uploads of the same vehicle and data from different vehicles, but cannot restore the association between the VIN and a specific vehicle owner, thus meeting the data anonymization requirements.
[0088] In this embodiment, the vehicle's location in the vehicle information can be converted into a regional location. The raw location accuracy of the Global Positioning System (GPS) can reach the meter level, and directly uploading it may leak privacy information. After conversion to a regional location, the cloud can still draw a map of premature combustion risk, but it cannot pinpoint the vehicle's specific activity trajectory, thus balancing privacy protection and functional requirements.
[0089] In this embodiment, personal information is not directly related to premature combustion risk analysis and strategy generation, and its removal does not affect the implementation of cloud functions. Removing personal information ensures that data collection meets the principle of minimum necessity, reducing the risk of data leakage and compliance costs.
[0090] In this embodiment, data anonymization forms a complete privacy protection process, ensuring that uploaded data cannot be associated with any specific vehicle during cloud storage, analysis, and sharing.
[0091] In this embodiment, the data uploaded to the cloud complies with domestic and international data security regulations (GDPR, Cybersecurity Law, and Data Security Law, etc.), and a user data authorization management interface is provided to respect users' right to know and right to choose.
[0092] In subsequent embodiments, the process of generating a pre-ignition risk map and an engine pre-ignition protection strategy in the cloud will be described.
[0093] As described above, the process involves irreversibly hashing the vehicle identification number (VIN), converting the vehicle location to a regional level, removing personal information from the vehicle information, and then anonymizing the processed vehicle information before sending it to the cloud. By performing irreversible hashing on the VIN, converting the location to a regional level, and anonymizing the personal information before uploading, the privacy leakage issue when uploading vehicle information from the vehicle to the cloud is resolved, thus balancing data security and collaborative analysis needs.
[0094] Based on the above embodiments, in order to achieve adaptive optimization throughout the entire life cycle, the engine protection strategy is continuously optimized in some embodiments.
[0095] Specifically, the process returns to the step of continuously collecting engine operating data, continuously collecting engine operating data and detecting pre-ignition risk. When a pre-ignition risk is detected, vehicle information and engine operating data corresponding to the pre-ignition risk are sent to the cloud so that the cloud can iteratively update the pre-ignition risk map, pre-ignition identification model and engine pre-ignition protection strategy.
[0096] In this embodiment, after performing pre-ignition protection on the engine, the system can return to the step of continuously collecting engine operating data. After pre-ignition protection is implemented, the engine combustion state changes, requiring continuous monitoring of the protection effect and potential side effects. Returning to the data collection step allows the system to maintain real-time awareness of the engine status, avoiding over- or under-protection measures.
[0097] For example, the iteration update cycle can be from 1 week to 4 weeks.
[0098] In this embodiment, the continuous monitoring and data feedback mechanism on the vehicle side after executing pre-ignition protection forms a complete closed loop of data collection → risk detection → strategy execution → effect feedback → strategy optimization. This enables the cloud-based pre-ignition risk map, pre-ignition identification model, and engine pre-ignition protection strategy to continuously iterate based on real vehicle operating data, adapting to dynamic factors such as engine aging, environmental changes, and new fuel formulas, thus achieving continuous closed-loop optimization between the vehicle side and the cloud.
[0099] Figure 3 A flowchart illustrating the engine pre-ignition control method provided in this application embodiment. Figure 3 The execution entity in this embodiment is the cloud. For example... Figure 3 As shown, the method includes: S301: Receives vehicle information and engine operating data sent by the vehicle terminal; the vehicle terminal continuously collects engine operating data; detects pre-ignition risk based on engine operating data; when a pre-ignition risk is detected, it sends vehicle information and the corresponding engine operating data to the cloud.
[0100] In this embodiment, the cloud receives and stores vehicle information and engine operation data uploaded by multiple vehicle terminals through a big data warehouse, with a data scale reaching millions of vehicles and billions of records. This massive amount of data enables the cloud to discover patterns that cannot be seen in single-vehicle data.
[0101] S302: Generate a pre-ignition risk map based on vehicle information from multiple vehicle terminals and engine operating data corresponding to pre-ignition risk.
[0102] In this embodiment, the pre-ignition risk map establishes a mapping relationship between multi-dimensional factors such as geospatial, temporal, and fuel quality and the probability of pre-ignition, revealing the modulating effect of environmental factors on pre-ignition.
[0103] Specifically, based on the regional location of each vehicle, the time when the risk of pre-ignition occurs, and the fuel quality corresponding to the regional location, the probability of pre-ignition under different regions, times, and fuel qualities is statistically analyzed to generate a pre-ignition risk map.
[0104] The vehicle information includes at least a regional location. Regional location protects individual vehicle privacy by preventing the pinpointing of specific trajectories, while providing sufficient spatial resolution to characterize the impact of environmental factors (altitude, temperature, etc., typically vary at a regional scale) on pre-ignition. Regional location discretizes the geographic space into several regional units, statistically analyzes the frequency of pre-ignition within each region, and maps the geographical distribution of pre-ignition risk.
[0105] The engine operation data corresponding to the pre-ignition risk includes at least the moment when the pre-ignition risk occurs. The moment of occurrence records the distribution of pre-ignition events along the timeline. The cloud extracts the corresponding time attributes such as date, season, and day / night cycle, and statistically analyzes the frequency of pre-ignition occurrences in different time periods. This time information can reveal seasonal patterns (such as increased pre-ignition risk during hot summer months) and diurnal variation patterns (such as higher risk shortly after a cold start), allowing for the plotting of the pre-ignition risk distribution over time.
[0106] The fuel quality corresponding to a regional location can be obtained by linking gas station information with a fuel testing database. The vehicle records the gas station location and refueling time, and this information is then linked to a publicly available third-party fuel testing database to query the corresponding gas station's fuel octane rating, sulfur content, aromatics content, and other quality indicators.
[0107] In this embodiment, the regional location, time, and fuel quality are used as three-dimensional indexes to calculate the probability of pre-ignition within each three-dimensional cell.
[0108] For example, the probability of pre-ignition can be expressed as the number of pre-ignition events per unit mileage, eliminating the impact of differences in vehicle mileage in different regions and making the risk levels in different regions and at different times comparable.
[0109] In this embodiment, the premature combustion risk map can be updated according to a preset period. Optionally, the preset period can be 24 hours.
[0110] S303: Based on engine operating data corresponding to pre-ignition risks from multiple vehicle terminals, obtain a trained pre-ignition recognition model.
[0111] Specifically, a federated learning framework is adopted to collaboratively train the pre-ignition identification model using engine operating data from multiple vehicles. During the collaborative training process, pre-ignition precursor features in the engine operating data are mined to obtain a global pre-ignition identification model, which is then used as the trained pre-ignition identification model.
[0112] In this embodiment, the federated learning framework enables collaborative model training while protecting data privacy.
[0113] For example, the federated learning framework may employ either the Federated Averaging (FedAvg) algorithm or the Federated Stochastic Gradient Descent (FedSGD) algorithm.
[0114] In this embodiment, the precursor features of pre-ignition are abnormal patterns that appear within dozens of combustion cycles before pre-ignition occurs, such as increased cylinder pressure fluctuation amplitude, enhanced high-frequency components of the vibration spectrum, and increased ignition delay. A deep learning model, by analyzing massive amounts of time-series data before and after pre-ignition risk, automatically learns the spatiotemporal evolution patterns of these precursor features. This eliminates the need for manually designed feature extraction algorithms and can uncover complex patterns that are difficult for engineers to summarize through experience.
[0115] In this embodiment, a federated learning framework is used to collaboratively train data from multiple vehicles, uncovering global early-ignition warning features and generating a global model with strong generalization ability and high robustness. On the one hand, it can identify complex early-ignition triggering mechanisms that are difficult to summarize using manual rules, significantly improving prediction accuracy. On the other hand, compared with models trained on a single vehicle, the accuracy is improved by more than 20%, and the false alarm rate is reduced by more than 50%.
[0116] S304: Generate the engine pre-ignition protection strategy corresponding to the current vehicle based on the pre-ignition risk map and pre-ignition identification model.
[0117] Specifically, the engine load and engine speed of the current vehicle are obtained; combined with the current regional location of the vehicle, the time when pre-ignition risk occurs, and the fuel quality corresponding to the regional location, the probability of pre-ignition is obtained from the pre-ignition risk map; based on the probability of pre-ignition, the ignition advance angle corresponding to the engine load and engine speed in the baseline ignition advance angle upper limit pulse spectrum table is corrected to obtain the ignition advance angle upper limit pulse spectrum table for the current vehicle; based on the probability of pre-ignition, the adjustment range of the engine control parameters for the current vehicle is determined; the ignition advance angle upper limit pulse spectrum table and the adjustment range of the engine control parameters are combined into engine control parameters; the model parameters of the pre-ignition identification model and the engine control parameters are used as the engine pre-ignition protection strategy for the current vehicle.
[0118] The upper limit pulse spectrum table of the reference ignition advance angle is a three-dimensional table, which includes engine load, engine speed and ignition advance angle.
[0119] In this embodiment, engine load and engine speed are used as lookup indexes to accurately locate the ignition advance angle allowed under the current operating conditions, thereby achieving fine-grained ignition angle correction based on actual operating conditions and avoiding unnecessary performance losses caused by uniformly adjusting the ignition advance angle for all operating conditions.
[0120] Next, the cloud will use the vehicle's current regional location, the moment when pre-ignition risk occurs, and the corresponding fuel quality at that regional location as the query keys for the pre-ignition risk map, retrieving the probability of pre-ignition under the corresponding conditions. This provides accurate risk quantification for determining the subsequent ignition advance angle correction range and control parameter adjustment range, ensuring that the strategy adjustment range matches the actual risk level.
[0121] Based on the pre-ignition probability obtained from the query, the ignition advance angle corresponding to engine load and engine speed in the baseline ignition advance angle upper limit pulse spectrum table is corrected to generate a personalized ignition advance angle upper limit pulse spectrum table adapted to the current vehicle usage environment.
[0122] In this embodiment, the ignition advance angle is dynamically adjusted based on the probability of pre-ignition risk. In areas with a high risk of pre-ignition (high-risk areas, low-octane fuel, high-temperature seasons), the upper limit of the ignition advance angle is reduced to improve the safety margin. When the risk is low, a higher ignition advance angle is maintained to balance power performance. Compared with a fixed ignition advance angle limitation strategy, this dynamic adjustment achieves a dynamic balance between safety and performance.
[0123] Meanwhile, based on the probability of pre-ignition risk, the cloud determines other engine control parameters besides the ignition advance angle, such as the engine torque limiting degree. The adjustable range for pre-ignition protection reflects the boundary of the control strategy's intervention force at the current risk level. By determining the probability-driven adjustment range, the intervention intensity of the control strategy is matched to the actual risk level. A larger adjustment range is set for high-risk scenarios to provide stronger protection; a smaller adjustment range is retained for low-risk scenarios to reduce the impact on power performance, avoiding the problem of fixed threshold strategies being overly conservative in low-risk situations and insufficiently protective in high-risk situations.
[0124] In this embodiment, the ignition advance angle upper limit pulse spectrum table and the adjustment range of engine control parameters are packaged into a complete set of engine control parameters, which serves as the control execution part of the engine pre-ignition protection strategy.
[0125] In this embodiment, the final engine pre-ignition protection strategy consists of two parts: model parameters of the pre-ignition detection model and engine control parameters. The model parameters of the pre-ignition detection model are used to update the local pre-ignition detection model on the vehicle; the engine control parameters are used to update the control strategy of the engine control unit. Together, they constitute a complete personalized engine pre-ignition protection strategy.
[0126] S305: Sends the engine pre-ignition protection strategy to the vehicle so that the vehicle can perform pre-ignition protection on the engine according to the engine pre-ignition protection strategy.
[0127] In this embodiment, a phased, regional gray-scale release mechanism can be adopted to distribute the engine pre-ignition protection strategy. Target areas are determined using a pre-ignition risk map, and a subset of vehicles within these areas are prioritized as the first batch of recipients. Engine pre-ignition protection strategies are matched to different vehicles based on their VINs and distributed via OTA. The cloud continuously receives engine operating data from the first batch of vehicles to evaluate the pre-ignition protection effect of the strategy. If no anomalies are detected, the batch of vehicles receiving the strategy is gradually expanded; if a risk is identified, the strategy distribution is paused.
[0128] In this embodiment, the OTA release interval can be 1-3 months, that is, the update frequency of the engine pre-ignition protection strategy is 1-3 months, which can be adjusted according to the actual situation.
[0129] As described above, the system receives vehicle information and engine operating data from the vehicle terminal. Specifically, the vehicle terminal continuously collects engine operating data; based on this data, it detects the risk of pre-ignition; upon detection, it sends vehicle information and the corresponding engine operating data to the cloud; a pre-ignition risk map is generated based on vehicle information and engine operating data from multiple vehicle terminals; a trained pre-ignition detection model is obtained based on this data; the pre-ignition risk map and the detection model generate a pre-ignition protection strategy for the current vehicle terminal; and this strategy is then sent to the vehicle terminal, enabling it to protect the engine from pre-ignition. By analyzing vehicle information and engine operating data from multiple vehicle terminals, generating a pre-ignition risk map and a trained pre-ignition detection model, the system can uncover complex patterns of pre-ignition within massive amounts of vehicle data and generate personalized engine protection strategies for the current vehicle terminal, thus improving the accuracy of pre-ignition protection.
[0130] Based on the above embodiments, in this embodiment, in order to further improve the reliability of the engine pre-ignition protection strategy before it is issued, a digital twin model of engine pre-ignition is constructed in the cloud to pre-verify the generated engine pre-ignition protection strategy.
[0131] Specifically, the engine pre-ignition protection strategy is input into the engine pre-ignition digital twin model, which then performs simulation verification of the engine pre-ignition protection strategy to obtain the pre-ignition risk prediction results corresponding to the engine pre-ignition protection strategy; based on the pre-ignition risk prediction results, the engine pre-ignition protection strategy is adjusted.
[0132] In this embodiment, the engine pre-ignition digital twin model can be a hybrid model combining a computational fluid dynamics (CFD) physical model with a data-driven model, integrating key factors affecting pre-ignition such as fuel characteristics, engine oil characteristics, and carbon deposit growth. The engine pre-ignition digital twin model provides the ability to perform rapid simulation verification without real-vehicle testing, significantly reducing strategy verification costs, while enabling the cloud to identify potential problems in the strategy before it is pushed out.
[0133] In this embodiment, the engine pre-ignition protection strategy to be verified is input into the engine pre-ignition digital twin model. Simulations are run under different operating conditions and environments, and the pre-ignition risk prediction results corresponding to the engine pre-ignition protection strategy are output. The pre-ignition risk prediction results generated by the simulation verification provide the cloud with a quantitative basis for evaluating the effectiveness of the engine pre-ignition protection strategy before real deployment. By simulating various extreme operating conditions, such as high load and low speed, low octane fuel, and high ambient temperature, the adequacy of the engine pre-ignition protection strategy under various conditions can be comprehensively evaluated, and potential risk scenarios that cannot be identified by static analysis can be discovered. This ensures the effectiveness and usability of the engine pre-ignition protection strategy sent to the vehicle, further improving the accuracy of pre-ignition protection.
[0134] Figure 4 A schematic diagram of the engine pre-ignition control device provided in the embodiments of this application. Figure 1 .like Figure 4 As shown, the engine pre-ignition control device, applied to the vehicle end, includes: a data acquisition module 401, a detection module 402, a first transmitting module 403, a first receiving module 404, and a pre-ignition protection module 405.
[0135] The acquisition module 401 is used to continuously acquire engine operating data.
[0136] The detection module 402 is used to detect the risk of pre-ignition based on engine operating data.
[0137] The first sending module 403 is used to send vehicle information and engine operating data corresponding to the pre-ignition risk to the cloud when a pre-ignition risk is detected, so that the cloud can generate a pre-ignition risk map based on the vehicle information and engine operating data corresponding to the pre-ignition risk from multiple vehicle terminals; obtain a trained pre-ignition recognition model based on the engine operating data corresponding to the pre-ignition risk from multiple vehicle terminals; and generate an engine pre-ignition protection strategy corresponding to the current vehicle terminal based on the pre-ignition risk map and the pre-ignition recognition model.
[0138] The first receiving module 404 is used to receive the engine pre-ignition protection strategy sent from the cloud.
[0139] The pre-ignition protection module 405 is used to protect the engine from pre-ignition according to the engine pre-ignition protection strategy.
[0140] In one possible implementation of this application, the acquisition module 401 is specifically used to: acquire engine operating data at a first sampling frequency when no risk of pre-ignition is detected; and switch to a second sampling frequency to acquire engine operating data when a risk of pre-ignition is detected; wherein the second sampling frequency is higher than the first sampling frequency.
[0141] In one possible implementation of this application, the first sending module 403 is specifically used to: based on the moment when a risk of pre-ignition is detected, extract engine operating data within a preset time window before the moment and engine operating data collected within a preset time window after the moment from the collected engine operating data; and send the engine operating data within the two preset time windows to the cloud.
[0142] In one possible implementation of this application, the first sending module 403 is further configured to: compress engine operating data within two preset time windows; and send the compressed engine operating data to the cloud during network idle periods.
[0143] In one possible implementation of this application, the vehicle information includes at least a vehicle identification number (VIN) and a vehicle location; the first sending module 403 is further configured to: perform irreversible hashing on the VIN in the vehicle information; convert the vehicle location in the vehicle information into a regional location; and remove personal information from the vehicle information; perform data anonymization on the processed vehicle information; and send the anonymized vehicle information to the cloud; wherein, the regional location is used by the cloud to collect engine operation data from multiple vehicle terminals in different regions and generate a pre-ignition risk map; the VIN is used for engine pre-ignition protection strategies generated for different vehicle terminals.
[0144] In one possible implementation of this application, the first receiving module 404 is specifically used for: receiving the engine pre-ignition protection strategy of the current vehicle end through an over-the-air download channel when the vehicle is in a shutdown and parking state; wherein the engine pre-ignition protection strategy carries a digital signature generated by the cloud for the engine pre-ignition protection strategy; verifying the digital signature of the engine pre-ignition protection strategy; and storing the engine pre-ignition protection strategy received this time after the digital signature verification is successful.
[0145] In one possible implementation of this application, the engine pre-ignition control device further includes an iterative update module. The iterative update module is specifically used to: return to the step of continuously collecting engine operating data, continuously collect engine operating data and detect pre-ignition risk, and when a pre-ignition risk is detected, send vehicle information and engine operating data corresponding to the pre-ignition risk to the cloud, so that the cloud iteratively updates the pre-ignition risk map, the pre-ignition identification model, and the engine pre-ignition protection strategy.
[0146] The apparatus provided in this application embodiment can be used to execute the technical solutions of the above method embodiments. Its implementation principle and technical effect are similar, and will not be repeated here.
[0147] Figure 5 A schematic diagram of the engine pre-ignition control device provided in the embodiments of this application. Figure 2 .like Figure 5 As shown, the engine pre-ignition control device is applied in the cloud and includes: a second receiving module 501, a first generating module 502, an acquisition module 503, a second generating module 504, and a second sending module 505.
[0148] The second receiving module 501 is used to receive vehicle information and engine operating data sent by the vehicle terminal; wherein, the vehicle terminal continuously collects engine operating data; detects the risk of pre-ignition based on the engine operating data; and when the risk of pre-ignition is detected, sends the vehicle information and the engine operating data corresponding to the pre-ignition risk to the cloud.
[0149] The first generation module 502 is used to generate a pre-ignition risk map based on vehicle information from multiple vehicle terminals and engine operating data corresponding to the pre-ignition risk.
[0150] The acquisition module 503 is used to acquire a trained pre-ignition recognition model based on engine operating data corresponding to pre-ignition risks from multiple vehicle terminals.
[0151] The second generation module 504 is used to generate the engine pre-ignition protection strategy corresponding to the current vehicle based on the pre-ignition risk map and the pre-ignition identification model.
[0152] The second sending module 505 is used to send the engine pre-ignition protection strategy to the vehicle end, so that the vehicle end can perform pre-ignition protection on the engine according to the engine pre-ignition protection strategy.
[0153] In one possible implementation of this application, the vehicle information includes at least a regional location; the engine operation data corresponding to the pre-ignition risk includes at least the time when the pre-ignition risk occurs; the first generation module 502 is specifically used to: based on the regional location of each vehicle, the time when the pre-ignition risk occurs, and the fuel quality corresponding to the regional location, to calculate the probability of pre-ignition occurring under different regions, different times, and different fuel qualities, and to generate a pre-ignition risk map.
[0154] In one possible implementation of this application, the acquisition module 503 is specifically used to: adopt a federated learning framework to collaboratively train the pre-ignition identification model using engine operating data from multiple vehicles; during the collaborative training process, mine the pre-ignition precursor features in the engine operating data to obtain a global pre-ignition identification model, and use the global pre-ignition identification model as the trained pre-ignition identification model.
[0155] In one possible implementation of this application, the second generation module 504 is specifically used for: obtaining the current engine load and engine speed of the vehicle; combining the current vehicle's regional location, the time when pre-ignition risk occurs, and the fuel quality corresponding to the regional location, to obtain the probability of pre-ignition from the pre-ignition risk map; based on the probability of pre-ignition, correcting the ignition advance angle corresponding to the engine load and engine speed in the baseline ignition advance angle upper limit pulse spectrum table to obtain the current vehicle's ignition advance angle upper limit pulse spectrum table; determining the adjustment range of the engine control parameters of the current vehicle based on the probability of pre-ignition; combining the ignition advance angle upper limit pulse spectrum table and the adjustment range of the engine control parameters into engine control parameters; and using the model parameters of the pre-ignition identification model and the engine control parameters as the engine pre-ignition protection strategy corresponding to the current vehicle.
[0156] In one possible implementation of this application, the engine pre-ignition control device further includes a simulation testing module. The simulation testing module is specifically used for: inputting the engine pre-ignition protection strategy into the engine pre-ignition digital twin model, enabling the engine pre-ignition digital twin model to simulate and verify the engine pre-ignition protection strategy, obtaining the pre-ignition risk prediction result corresponding to the engine pre-ignition protection strategy; and adjusting the engine pre-ignition protection strategy based on the pre-ignition risk prediction result.
[0157] The apparatus provided in this application embodiment can be used to execute the technical solutions of the above method embodiments. Its implementation principle and technical effect are similar, and will not be repeated here.
[0158] Figure 6 This is a schematic diagram of the vehicle-side hardware structure provided in an embodiment of this application. Figure 6 As shown, the vehicle end of this embodiment includes a processor 601 and a memory 602.
[0159] The memory 602 stores computer-executed instructions; the processor 601 executes the computer-executed instructions stored in the memory to implement the various steps performed by the vehicle in the above embodiments. For details, please refer to the relevant descriptions in the foregoing method embodiments.
[0160] Alternatively, the memory 602 can be either standalone or integrated with the processor 601.
[0161] When the memory 602 is set up independently, the vehicle end also includes a bus 603 for connecting the memory 602 and the processor 601.
[0162] Figure 7 This is a schematic diagram of the cloud hardware structure provided in an embodiment of this application. Figure 7 As shown, the cloud in this embodiment includes a processor 701 and a memory 702.
[0163] The memory 702 stores computer execution instructions; the processor 701 executes the computer execution instructions stored in the memory to implement the various steps performed in the cloud in the above embodiments. For details, please refer to the relevant descriptions in the foregoing method embodiments.
[0164] Alternatively, the memory 702 can be either standalone or integrated with the processor 701.
[0165] When the memory 702 is set up independently, the cloud also includes a bus 703 for connecting the memory 702 and the processor 701.
[0166] This application also provides a computer-readable storage medium storing computer-executable instructions. When a processor executes the computer-executable instructions, the above-described engine pre-ignition control method is implemented.
[0167] This application also provides a computer program product, including a computer program, which, when executed by a processor, implements the above-described engine pre-ignition control method.
[0168] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or modules, and may be electrical, mechanical, or other forms.
[0169] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to implement the solution of this embodiment according to actual needs.
[0170] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The unit composed of the above modules can be implemented in hardware or in the form of hardware plus software functional units.
[0171] The integrated modules described above, implemented as software functional modules, can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods of the various embodiments of this application.
[0172] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. A general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly manifested as execution by a hardware processor, or execution by a combination of hardware and software modules within the processor.
[0173] The memory may include high-speed RAM, and may also include non-volatile storage (NVM), such as at least one disk storage device, and may also be a USB flash drive, external hard drive, read-only memory, disk or optical disc, etc.
[0174] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.
[0175] The aforementioned storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The storage medium can be any available medium that can be accessed by a general-purpose or special-purpose computer.
[0176] An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Alternatively, the storage medium can be an integral part of the processor. Both the processor and the storage medium can reside in an Application Specific Integrated Circuit (ASIC). Alternatively, the processor and storage medium can exist as discrete components in an electronic device or host device.
[0177] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.
[0178] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.
Claims
1. A method for controlling engine pre-ignition, characterized in that, Applications in vehicles include: Continuously collect engine operating data; Based on the engine operating data, detect whether there is a risk of pre-ignition; When the risk of pre-ignition is detected, vehicle information and engine operating data corresponding to the pre-ignition risk are sent to the cloud. The cloud then generates a pre-ignition risk map based on the vehicle information and engine operating data from multiple vehicle terminals. Based on the engine operating data from multiple vehicle terminals, a trained pre-ignition recognition model is obtained. Based on the pre-ignition risk map and the pre-ignition recognition model, a pre-ignition protection strategy for the engine corresponding to the current vehicle terminal is generated. Receive the engine pre-ignition protection strategy sent from the cloud; According to the engine pre-ignition protection strategy, the engine is protected against pre-ignition.
2. The method according to claim 1, characterized in that, The continuous collection of engine operating data includes: When no pre-ignition risk is detected, engine operating data is collected at a first sampling frequency; When the risk of pre-ignition is detected, the engine operation data is collected at a second sampling frequency, wherein the second sampling frequency is higher than the first sampling frequency.
3. The method according to claim 1, characterized in that, Send the engine operating data corresponding to the aforementioned pre-ignition risk to the cloud, including: Based on the moment when the risk of pre-ignition is detected, engine operating data within a preset time window before the detected moment and engine operating data within a preset time window after the detected moment are extracted from the collected engine operating data. The engine operation data within two preset time windows is sent to the cloud.
4. The method according to claim 3, characterized in that, The engine operation data within the two preset time windows is sent to the cloud, including: Compress engine operating data within two preset time windows; During periods of network downtime, compressed engine operating data is sent to the cloud.
5. The method according to claim 1, characterized in that, The vehicle information includes at least the vehicle identification number and the vehicle location; Sending the vehicle information to the cloud includes: The vehicle identification number in the vehicle information is subjected to irreversible hashing; the vehicle location in the vehicle information is converted into a regional location; and personal information in the vehicle information is removed. The processed vehicle information is anonymized, and the anonymized vehicle information is sent to the cloud. The regional location is used by the cloud to collect engine operation data from multiple vehicles and generate the pre-ignition risk map; the vehicle identification number is used for engine pre-ignition protection strategies generated for different vehicles.
6. The method according to claim 1, characterized in that, The process of receiving the engine pre-ignition protection strategy sent from the cloud includes: When the vehicle is in a stopped state with the engine off, the engine pre-ignition protection strategy is received from the vehicle via an over-the-air download channel; wherein the engine pre-ignition protection strategy carries a digital signature generated by the cloud for the engine pre-ignition protection strategy. The engine pre-ignition protection strategy is digitally verified; After the digital signature verification is successful, the engine pre-ignition protection strategy received this time is stored.
7. The method according to any one of claims 1-6, characterized in that, After implementing pre-ignition protection for the engine, the following is also included: Return to the step of continuously collecting engine operating data, continuously collect the engine operating data and detect the risk of pre-ignition. When the risk of pre-ignition is detected, send the vehicle information and the engine operating data corresponding to the pre-ignition risk to the cloud so that the cloud can iteratively update the pre-ignition risk map, the pre-ignition identification model and the engine pre-ignition protection strategy.
8. A method for controlling engine pre-ignition, characterized in that, Applied to the cloud, including: The system receives vehicle information and engine operating data sent from the vehicle terminal; wherein the vehicle terminal continuously collects engine operating data; based on the engine operating data, it detects the risk of pre-ignition; when the risk of pre-ignition is detected, it sends the vehicle information and the engine operating data corresponding to the pre-ignition risk to the cloud. A pre-ignition risk map is generated based on the vehicle information from multiple vehicle terminals and the engine operating data corresponding to the pre-ignition risk. Based on engine operating data corresponding to the pre-ignition risk from multiple vehicle terminals, a trained pre-ignition identification model is obtained; Based on the pre-ignition risk map and the pre-ignition identification model, generate the engine pre-ignition protection strategy corresponding to the current vehicle. The engine pre-ignition protection strategy is sent to the vehicle, so that the vehicle can perform pre-ignition protection on the engine according to the engine pre-ignition protection strategy.
9. The method according to claim 8, characterized in that, The vehicle information includes at least a regional location; the engine operating data corresponding to the pre-ignition risk includes at least the time when the pre-ignition risk occurred. The step of generating a pre-ignition risk map based on vehicle information from multiple vehicle terminals and engine operating data corresponding to the pre-ignition risk includes: Based on the regional location of each vehicle, the time when the risk of pre-ignition occurs, and the fuel quality corresponding to the regional location, the probability of pre-ignition under different regions, times, and fuel qualities is statistically analyzed to generate a pre-ignition risk map.
10. The method according to claim 8, characterized in that, The step of obtaining a trained pre-ignition identification model based on engine operating data corresponding to the pre-ignition risk from multiple vehicle terminals includes: A federated learning framework is adopted to collaboratively train the pre-ignition identification model using engine operation data from multiple vehicles. During the collaborative training process, the pre-ignition precursor features in the engine operating data are mined to obtain a global pre-ignition identification model, and the global pre-ignition identification model is used as the trained pre-ignition identification model.
11. The method according to claim 8, characterized in that, The step of generating an engine pre-ignition protection strategy for the current vehicle based on the pre-ignition risk map and the pre-ignition identification model includes: Obtain the current engine load and engine speed on the vehicle. By combining the current regional location of the vehicle, the time when the risk of pre-ignition occurs, and the fuel quality corresponding to the regional location, the probability of pre-ignition is obtained from the pre-ignition risk map. Based on the probability of pre-ignition, the ignition advance angle corresponding to the engine load and the engine speed in the reference ignition advance angle upper limit pulse spectrum table is corrected to obtain the current vehicle ignition advance angle upper limit pulse spectrum table. Based on the probability of pre-ignition, determine the adjustment range of the engine control parameters at the current vehicle end; The upper limit pulse spectrum of the ignition advance angle and the adjustment range of the engine control parameters are combined to form the engine control parameters; The model parameters of the pre-ignition identification model and the engine control parameters are used as the engine pre-ignition protection strategy corresponding to the current vehicle.
12. The method according to any one of claims 8-11, characterized in that, Before sending the engine pre-ignition protection strategy to the vehicle, the method further includes: The engine pre-ignition protection strategy is input into the engine pre-ignition digital twin model, so that the engine pre-ignition digital twin model can perform simulation verification of the engine pre-ignition protection strategy and obtain the pre-ignition risk prediction result corresponding to the engine pre-ignition protection strategy. Based on the predicted pre-ignition risk, the engine pre-ignition protection strategy is adjusted.
13. A vehicle end, characterized in that, include: At least one processor; The at least one processor is also connected in communication with a memory, wherein the memory stores a computer program that can be executed by the at least one processor to cause the at least one processor to perform the engine pre-ignition control method according to any one of claims 1 to 7.
14. A cloud platform, characterized in that, include: At least one processor; And a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, the computer program being executed by the at least one processor to cause the at least one processor to perform the engine pre-ignition control method according to any one of claims 8 to 12.
15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the engine pre-ignition control method as described in any one of claims 1 to 7 or 8 to 12.
16. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the engine pre-ignition control method as described in any one of claims 1 to 7 or 8 to 12.
