A method for controlling the rotational speed of a brush motor

Abstract

The application provides a brush motor rotating speed control method, relates to the technical field of data processing, and comprises the following steps: collecting the voltage, current and rotating speed of the brush motor, calculating the change rates of the voltage, current and rotating speed at adjacent time points, respectively constructing corresponding change rate sequences, dividing the change rate sequences into multiple window intervals, calculating delay amount groups in each window interval, judging whether the brush motor has response mutation, recording corresponding response abnormal time points, extracting the speed evolution data corresponding to each response abnormal time point, calculating the weighted average rotating speed change rate, and obtaining a predicted rotating speed offset sequence; according to the difference between the predicted rotating speed offset sequence and the real-time rotating speed, the compensation voltage increment is calculated, the compensation voltage signal is generated, and the rotating speed of the motor is ensured to be stable. The application can predict the rotating speed of the brush motor and correct the rotating speed through voltage to ensure the rotating speed to be stable.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method for controlling the speed of a brushed motor. Background Technology

[0002] In existing technologies, brushed motors are widely used in portable power tools, such as electric screwdrivers and electric drills, due to their simple structure and convenient control. These tools typically use voltage regulation to control the motor speed, for example, by indirectly controlling the voltage applied to the motor through thyristor voltage regulation or pulse width modulation signals, thereby adjusting the speed. This control method is inexpensive, has relatively simple control logic, and is suitable for applications where high precision is not required.

[0003] However, in environments where power tools are frequently started and stopped or experience large load changes, traditional voltage control methods may have shortcomings when dealing with sudden load changes. For example, when a power tool comes into contact with a hard material and the load suddenly increases, the speed control system may lag behind and fail to adjust the voltage compensation in time, causing the motor speed to drop instantly and resulting in a jamming phenomenon. Summary of the Invention

[0004] The purpose of this invention is to provide a method for controlling the speed of a brushed motor, which aims to solve the problems mentioned in the background art.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0006] A method for controlling the speed of a brushed motor, the method comprising:

[0007] The voltage, current, and speed of the brushed motor were collected and time-aligned to obtain a ternary synchronization dataset.

[0008] Based on the three-dimensional synchronization dataset, the rate of change of voltage, current and rotation speed at adjacent time points is calculated, and the corresponding rate of change sequences are constructed to obtain the rate evolution dataset.

[0009] Based on a preset time window, the rate of change sequence in the rate evolution dataset is divided into multiple window intervals, and within each window interval, the delay quantity group is calculated to obtain the response delay sequence.

[0010] By calculating the difference between the response delay sequence and the preset delay group, it is determined whether there is a sudden change in the response of the brushed motor, and the corresponding abnormal response time is recorded to obtain the change identification sequence.

[0011] Based on the mutation identification sequence, the rate evolution data corresponding to each abnormal response moment is extracted, the weighted average speed change rate is calculated, and the predicted speed offset sequence is obtained.

[0012] The compensation voltage increment is calculated based on the difference between the predicted speed offset sequence and the real-time speed to obtain the corrected voltage data;

[0013] Based on the corrected voltage data, a compensation voltage signal is generated and transmitted to the drive input of the brushed motor to ensure stable motor speed.

[0014] Furthermore, based on the ternary synchronization dataset, the rates of change of voltage, current, and rotational speed at adjacent time points are calculated, and corresponding rate of change sequences are constructed to obtain a rate evolution dataset, including:

[0015] Based on the difference between the current voltage value and the previous voltage value in the three-dimensional synchronous dataset, and combined with the sampling interval time, the voltage change rate is calculated to obtain the voltage change rate sequence.

[0016] Based on the difference between the current value at the current moment and the current value at the previous moment in the three-dimensional synchronous dataset, and combined with the sampling interval time, the current change rate is calculated to obtain the current change rate sequence.

[0017] Based on the difference between the current rotational speed value and the previous rotational speed value in the three-dimensional synchronous dataset, and combined with the sampling interval time, the rotational speed change rate is calculated to obtain the rotational speed change rate sequence.

[0018] A rate evolution dataset is constructed based on the voltage change rate sequence, current change rate sequence, and rotational speed change rate sequence.

[0019] Furthermore, based on a preset time window, the rate of change sequence in the rate evolution dataset is divided into multiple window intervals, and within each window interval, a delay quantity group is calculated to obtain the response delay sequence, including:

[0020] Based on the preset time window length, the voltage change rate sequence, current change rate sequence, and speed change rate sequence are divided into multiple window intervals according to time sequence to obtain multiple rate windows;

[0021] Within each rate window, the corresponding voltage change rate sequence and speed change rate sequence are extracted, sequence deviation matching is performed, the minimum delay time of speed relative to voltage is calculated, and the voltage delay value is obtained.

[0022] Within the same rate window, the current change rate sequence is extracted, the same sequence deviation matching is performed, the minimum delay time of rotational speed relative to current is calculated, and the current delay value is obtained.

[0023] Based on the voltage delay value and current delay value within each rate window, a delay value group is formed, and these groups are arranged in time order to obtain the response delay sequence.

[0024] Furthermore, within each rate window, the corresponding voltage change rate sequence and speed change rate sequence are extracted, sequence deviation matching is performed, and the minimum delay time of speed relative to voltage is calculated to obtain the voltage delay value, including:

[0025] Based on the start and end times of the rate window, continuous data points of the voltage change rate sequence and the speed change rate sequence within the rate window are extracted to obtain voltage change rate window segments and speed change rate window segments.

[0026] Keeping the position of the voltage change rate window segment unchanged, slide the speed change rate window segment backward along the time axis within the current rate window by several sampling points to construct the first speed delay sequence;

[0027] The voltage change rate window segment and the first rotational speed delay sequence are normalized and their numerical range is scaled to the [0,1] interval to obtain the first normalized segment set;

[0028] Based on the first normalized fragment set, the average absolute error between the voltage and its corresponding rotational speed is calculated to obtain the voltage slip offset value data.

[0029] Based on the voltage sliding offset data, the number of sliding sampling points corresponding to the minimum voltage sliding offset value is identified. Combined with the sampling interval time, the offset time of the rotational speed relative to the voltage within the current rate window is determined, and the voltage delay value is obtained.

[0030] Furthermore, within the same rate window, the current change rate sequence is extracted, and the same sequence deviation matching is performed. The minimum delay time of rotational speed relative to current is calculated to obtain the current delay value, including:

[0031] Within the same rate window, extract continuous data points of the current change rate sequence within that rate window to obtain a current change rate window segment.

[0032] Keeping the position of the current rate of change window segment unchanged, the rotational speed rate of change window segment is slid backward along the time axis by several sampling points within the current rate window to construct a second rotational speed delay sequence;

[0033] The current change rate window segment and the second rotational speed delay sequence are normalized to scale their numerical range to the [0,1] interval, thus obtaining the second normalized segment set;

[0034] Based on the second normalized fragment set, the average absolute error between the current and its corresponding rotational speed is calculated to obtain the current slip offset value data.

[0035] Based on the current sliding offset value data, the number of sliding sampling points corresponding to the minimum current sliding offset value is identified. Combined with the sampling interval time, the offset time of the rotation speed relative to the current within the current rate window is determined, and the current delay value is obtained.

[0036] Furthermore, by calculating the difference between the response delay sequence and the preset delay group, it is determined whether there is a sudden change in the response of the brushed motor, and the corresponding abnormal response time is recorded to obtain the change identification sequence, including:

[0037] Based on the voltage delay value and current delay value of each rate window in the response delay sequence, the difference between them and the preset voltage delay reference value and the preset current delay reference value is calculated to obtain the voltage delay difference and the current delay difference.

[0038] The absolute values ​​of the voltage delay difference and the current delay difference are processed to obtain non-negative delay offset pairs;

[0039] Determine whether any component in the non-negative delay offset value pair exceeds the preset offset threshold. If the result is yes, then there is a sudden change in response within the rate window, and the start time of the rate window is marked as the moment of abnormal response.

[0040] Arrange all abnormal response moments in chronological order to obtain the mutation identification sequence.

[0041] Furthermore, based on the mutation identification sequence, the rate evolution data corresponding to each abnormal response moment is extracted, and the weighted average rotational speed change rate is calculated to obtain the predicted rotational speed offset sequence, including:

[0042] Based on the abnormal response moments in the mutation identification sequence, the rotational speed change rate, current change rate, and voltage change rate data within a fixed prediction time starting from any abnormal response moment are extracted from the rate evolution dataset to obtain the abnormal response dataset.

[0043] Based on the response anomaly dataset, the weighting factors of the current change rate and voltage change rate are calculated, and combined with the speed change rate data, the weighted average speed change rate at the moment of the response anomaly is calculated.

[0044] The predicted speed offset is obtained based on the weighted average speed change rate and a fixed prediction duration.

[0045] Arrange the predicted speed offsets corresponding to all abnormal response times in chronological order to obtain the predicted speed offset sequence.

[0046] Furthermore, based on the response anomaly dataset, weighting factors for the current and voltage change rates are calculated, and combined with the speed change rate data, the weighted average speed change rate at the moment of the response anomaly is calculated, including:

[0047] Within a fixed prediction time at any time of response anomaly, the rate of change of current at each sampling point in the response anomaly segment is read sequentially, and the current weight value is obtained based on its absolute value and the preset current adjustment value.

[0048] The voltage change rate of each sampling point in the same response anomaly segment is read sequentially, and the voltage weight value is obtained based on its absolute value and the preset voltage adjustment value.

[0049] The current weight value, voltage weight value, and preset constant adjustment value corresponding to each sampling point are fused together to obtain a comprehensive weighting factor;

[0050] The weighted rate of change of rotational speed is obtained by multiplying the rate of change of rotational speed at each sampling point with its corresponding comprehensive weighting factor.

[0051] The weighted average speed change rate is obtained by summing all weighted speed change rates within the abnormal response segment and dividing it by the sum of all comprehensive weighting factors within the abnormal response segment.

[0052] Furthermore, based on the difference between the predicted speed offset sequence and the real-time speed, the compensation voltage increment is calculated to obtain the corrected voltage data, including:

[0053] Based on the predicted speed offset sequence, extract the predicted speed offset at any time of response anomaly, and extract the real-time speed at the current time.

[0054] The difference between the predicted speed offset and the real-time speed is calculated to obtain the speed compensation requirement value.

[0055] Within a fixed prediction time at the moment of the response anomaly, the voltage change rate and the speed change rate are extracted, and the average corresponding ratio between the two is calculated to obtain the response conversion factor.

[0056] Calculate the compensation voltage increment based on the response conversion factor and the speed compensation requirement;

[0057] The corrected voltage data is obtained based on the compensation voltage increment and the control voltage at the current moment.

[0058] Furthermore, based on the corrected voltage data, a compensation voltage signal is generated and transmitted to the drive input of the brushed motor to ensure stable motor speed, including:

[0059] Calculate the change in the corrected voltage data compared to the control voltage at the previous moment to obtain the correction magnitude;

[0060] If the correction magnitude exceeds the preset voltage adjustment threshold, the correction magnitude will be limited to ensure that the correction magnitude does not exceed the preset voltage adjustment threshold.

[0061] Determine whether the corrected voltage data exceeds the brushed motor voltage output range. If the result is yes, then limit the corrected voltage data to the brushed motor voltage output range.

[0062] After the above restrictions, the restricted correction voltage data is obtained. The weighted average of the restricted correction voltage data and the corrected voltage data is then used to obtain the smoothed correction voltage data.

[0063] Based on the smoothed correction voltage data, its value is converted into a voltage output command, a compensation voltage signal is generated, and it is transmitted to the drive input terminal of the brushed motor.

[0064] The above-described solution of the present invention has at least the following beneficial effects:

[0065] This invention transforms raw synchronization data into a rate-of-change sequence of voltage, current, and rotational speed, forming a rate evolution dataset. This is equivalent to introducing the dynamic expressive power of multivariable first-order difference features at the data processing level. Compared to traditional static sampling values, this evolution sequence can more realistically reflect the trend changes, abrupt changes, and inertial response characteristics of the system state. Through rate-of-change modeling, the motor system is transformed from point sampling to continuous trajectory modeling. Furthermore, this rate-of-change modeling process is more resistant to noise interference, facilitating the construction of predictive models using windowing mechanisms, and providing an accurate dynamic data foundation for the system.

[0066] This invention achieves refined modeling of the short-term operating state of the system by dividing the rate evolution dataset into time windows and analyzing the local response characteristics in each window. The feature parameters within each time window constitute a complete response unit, thereby avoiding the dilution problem of local abnormal behavior under the global feature extraction method. This processing mechanism has obvious data segmentation and sliding sampling characteristics, which significantly enhances the system's ability to detect local load changes. By quantifying and modeling the local response offset, it is possible to quickly calibrate the abrupt behavior and accurately lock the target window, providing a clear boundary range and data entry for predictive control.

[0067] This invention acquires response delay sequences and performs difference analysis with preset delay groups to identify abrupt change windows, constructing a dynamic coupling relationship identification mechanism for voltage, current changes and speed response. It realizes a full-chain data processing path from feature alignment to anomaly detection. This delay identification mechanism essentially uses time correlation to establish a lag matching relationship between input and output, avoiding false alarms or missed detections caused by single-variable fluctuation judgment. It can quickly pinpoint the moment when unwanted disturbance behavior occurs, providing highly timely criteria for predictive control.

[0068] This invention extracts evolution data for a specific time period by identifying mutation sequences, constructs a weighted average speed change rate, predicts future speed deviation trends, and generates a predicted speed deviation sequence. This mechanism integrates multiple modeling methods such as parameter collaborative analysis, voltage and current weight allocation, and time window integration, making the prediction results physically based, trend reasonable, and dynamically adaptable. This prediction mechanism can not only be used for the design of compensation strategies for sudden load disturbance responses, but also generate correction parameters in advance in persistent deviation trends, thereby enhancing the stability of the control system.

[0069] This invention collects three types of operating parameters of a brushed motor—voltage, current, and speed—and aligns them in the time dimension to form a high-time-resolution, highly consistent ternary synchronous dataset. This provides a fundamental guarantee for subsequent dynamic behavior modeling. This synchronous acquisition-alignment mechanism is significantly superior to the isolated sampling of parameters in existing technologies. It can effectively suppress data misalignment problems caused by sensor delays, jitter, or interruptions. This structured data model can ensure the traceability and real-time performance of analysis and calculation, providing a stable data foundation for subsequent calculation stages. Attached Figure Description

[0070] Figure 1 This is a flowchart of a brushed motor speed control method provided by an embodiment of the present invention. Detailed Implementation

[0071] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0072] like Figure 1 As shown, an embodiment of the present invention proposes a method for controlling the speed of a brushed motor, the method comprising:

[0073] The voltage, current, and speed of the brushed motor were collected and time-aligned to obtain a ternary synchronization dataset.

[0074] Based on the three-dimensional synchronization dataset, the rate of change of voltage, current and rotation speed at adjacent time points is calculated, and the corresponding rate of change sequences are constructed to obtain the rate evolution dataset.

[0075] Based on a preset time window, the rate of change sequence in the rate evolution dataset is divided into multiple window intervals, and within each window interval, the delay quantity group is calculated to obtain the response delay sequence.

[0076] By calculating the difference between the response delay sequence and the preset delay group, it is determined whether there is a sudden change in the response of the brushed motor, and the corresponding abnormal response time is recorded to obtain the change identification sequence.

[0077] Based on the mutation identification sequence, the rate evolution data corresponding to each abnormal response moment is extracted, the weighted average speed change rate is calculated, and the predicted speed offset sequence is obtained.

[0078] The compensation voltage increment is calculated based on the difference between the predicted speed offset sequence and the real-time speed to obtain the corrected voltage data;

[0079] Based on the corrected voltage data, a compensation voltage signal is generated and transmitted to the drive input of the brushed motor to ensure stable motor speed.

[0080] In this embodiment of the invention, the voltage, current, and speed of the brushed motor are collected and time-aligned to obtain a ternary synchronous dataset. This ensures that the three types of key operating data have time consistency and source synchronization, avoiding calculation errors or confusion of causal relationships caused by time offsets in subsequent processing, and providing accurate and stable raw data support for subsequent processing. Based on the ternary synchronous dataset, the rate of change of voltage, current, and speed at adjacent moments are calculated, and corresponding rate of change sequences are constructed to obtain a rate evolution dataset. This captures the response speed and dramatic behavior of the motor's operating state at a microscopic time scale, improving the sensitivity to sudden load disturbances and the ability to make dynamic judgments. According to a preset time window, the rate of change sequence in the rate evolution dataset is divided into multiple window intervals, and within each window interval, delay groups are calculated to obtain a response delay sequence. This constructs a hysteresis mapping relationship between the motor input and output, effectively identifying the response offset trend within a local time period, which helps to quantify the degree of dynamic coupling from a time dimension.

[0081] By calculating the difference between the response delay sequence and the preset delay group, it is determined whether there is a sudden change in the response of the brushed motor, and the corresponding abnormal response time is recorded to obtain the change identification sequence. The statistical deviation of the response lag is judged to achieve early capture and local time domain locking of system operation anomalies, and accurately screen out potential fault or load fluctuation risk periods. According to the change identification sequence, the rate evolution data corresponding to each abnormal response time is extracted, the weighted average speed change rate is calculated, and the predicted speed offset sequence is obtained. This enables advanced prediction of the motor operation trend under abnormal conditions, generates short-term control target quantities, and improves the initiative and advanced correction capability of the control system. According to the difference between the predicted speed offset sequence and the real-time speed, the compensation voltage increment is calculated to obtain the correction voltage data, realizing the mapping from the predicted control quantity to the executed voltage quantity, ensuring that the voltage compensation has a quantitative basis, calculation consistency and dynamic adaptability. According to the correction voltage data, the compensation voltage signal is generated and transmitted to the drive input of the brushed motor to ensure the stability of the motor speed. On the basis of ensuring the adjustment effect, the stability and safety of the control system are enhanced, providing a good terminal guarantee for achieving precise and flexible speed regulation.

[0082] This involves collecting the voltage, current, and speed of a brushed motor, and then aligning them in time to obtain a ternary synchronization dataset, specifically including:

[0083] First, sensing devices for real-time measurement of voltage, current, and speed need to be deployed in the motor system. Voltage sampling can be performed using a voltage divider circuit combined with a high-precision analog-to-digital converter (ADC) to convert the instantaneous voltage between the two poles of the motor terminals into a digital signal. Current acquisition typically uses Hall current sensors or sampling resistors to convert changes in current in the motor power supply line into a voltage signal in real time, which is then input to the ADC for digitization. The method for acquiring the speed signal varies depending on the motor structure. In general portable power tools, photoelectric encoders, Hall position sensors, or sensorless algorithms based on back EMF waveform analysis can be used to estimate the speed. The output value needs to be further smoothed by a filtering module to remove high-speed fluctuation noise, ultimately forming a speed sample value.

[0084] The signal sources for the three physical quantities—voltage, current, and rotational speed—are different, resulting in varying delay characteristics in their sampling links, which can easily lead to asynchronous data acquisition on the time axis. To address this, a unified high-precision clock is used as the sampling trigger reference for the main control unit. This allows for unified scheduling of the three sampling modules, and parallel reading of each data channel is achieved using DMA, avoiding time jitter caused by CPU resource switching. In actual data acquisition, the system triggers a synchronous reading operation every set of sampling periods (e.g., every 1 millisecond) and assigns a unique timestamp to each set of sampled values ​​to ensure strict consistency of the three data streams on the time stamp.

[0085] To further improve the accuracy of time alignment, linear interpolation or spline interpolation algorithms can be used to adjust samples that are slightly out of sync due to hardware delays. Specifically, using a certain data stream as a reference, its nearest neighbor value on the time axis is interpolated on a linear scale to estimate the value that the other data stream should have at the corresponding time point, thus forming paired data samples of the three quantities at the same time point. After the alignment operation is completed, the voltage, current, and rotational speed values ​​at each time point can be combined into a ternary vector data item, forming a ternary synchronization data sequence. The set of consecutive sample points constitutes the ternary synchronization dataset.

[0086] In a preferred embodiment of the present invention, based on a ternary synchronization dataset, the rates of change of voltage, current, and rotational speed at adjacent time points are calculated, and corresponding rate of change sequences are constructed to obtain a rate evolution dataset, including:

[0087] Based on the difference between the current voltage value and the previous voltage value in the three-dimensional synchronous dataset, and combined with the sampling interval time, the voltage change rate is calculated to obtain the voltage change rate sequence.

[0088] Based on the difference between the current value at the current moment and the current value at the previous moment in the three-dimensional synchronous dataset, and combined with the sampling interval time, the current change rate is calculated to obtain the current change rate sequence.

[0089] Based on the difference between the current rotational speed value and the previous rotational speed value in the three-dimensional synchronous dataset, and combined with the sampling interval time, the rotational speed change rate is calculated to obtain the rotational speed change rate sequence.

[0090] A rate evolution dataset is constructed based on the voltage change rate sequence, current change rate sequence, and rotational speed change rate sequence.

[0091] In this embodiment of the invention, the voltage change rate is calculated based on the difference between the current voltage value and the previous voltage value in the ternary synchronous dataset, combined with the sampling interval time, to obtain a voltage change rate sequence. This extracts the voltage change trend on the time axis, accurately describing the actual input intensity of the control signal to the motor's dynamic response, and providing a more dynamic data foundation for subsequent delay matching and abrupt change identification. Similarly, the current change rate is calculated based on the difference between the current value and the previous current value in the ternary synchronous dataset, combined with the sampling interval time, to obtain a current change rate sequence, determining whether the motor is in a normal acceleration, overload, or abrupt change state. Furthermore, the speed change rate is calculated based on the difference between the current speed value and the previous speed value in the ternary synchronous dataset, combined with the sampling interval time, to obtain a speed change rate sequence. By comparing this with the voltage change rate and current change rate, the response efficiency and hysteresis characteristics between the system input and output are derived. Finally, the voltage change rate sequence, current change rate sequence, and speed change rate sequence form a rate evolution dataset, effectively enhancing the ability to analyze the linkage between signals and laying a data foundation for dynamic correlation modeling and response trend reasoning.

[0092] Specifically, based on the difference between the current voltage value and the previous voltage value in the ternary synchronous dataset, and combined with the sampling interval, the voltage change rate is calculated to obtain the voltage change rate sequence, which includes:

[0093] First, a voltage signal sequence is extracted from the obtained ternary synchronous dataset. Data sampling is performed at fixed time intervals to ensure that there is an equidistant time interval between any two adjacent sampling points in the sequence. Specifically, the sampling time interval is set to... (Unit: seconds), then the voltage sequence can be denoted as: ,in The voltage value at the initial moment. This is the voltage value for the next sampling period, and so on.

[0094] Subsequently, for each pair of adjacent sampling points in the voltage sequence, a difference calculation operation is performed. (The sentence is incomplete and requires further context.) time( For example, its corresponding current voltage value is The voltage value at the previous moment was The difference between the two reflects the change in voltage per unit sampling time. To normalize this change to a unit time, the following formula is used for normalization: ,in Indicates the first The rate of change of voltage at any given time (in V / s). and These are the voltage values ​​at the current and previous sampling points, respectively. This is the preset sampling interval.

[0095] Following the above calculation method, by traversing the entire voltage sampling sequence one by one, a length of [length missing] can be obtained. Voltage change rate sequence Each term represents the voltage dynamic change trend within the corresponding time period, and the resulting voltage change rate sequence will be used as part of the rate evolution dataset for subsequent steps.

[0096] In a preferred embodiment of the present invention, the rate of change sequence in the rate evolution dataset is divided into multiple window intervals according to a preset time window, and a delay quantity group is calculated within each window interval to obtain a response delay sequence, including:

[0097] Based on the preset time window length, the voltage change rate sequence, current change rate sequence, and speed change rate sequence are divided into multiple window intervals according to time sequence to obtain multiple rate windows;

[0098] Within each rate window, the corresponding voltage change rate sequence and speed change rate sequence are extracted, sequence deviation matching is performed, the minimum delay time of speed relative to voltage is calculated, and the voltage delay value is obtained.

[0099] Within the same rate window, the current change rate sequence is extracted, the same sequence deviation matching is performed, the minimum delay time of rotational speed relative to current is calculated, and the current delay value is obtained.

[0100] Based on the voltage delay value and current delay value within each rate window, a delay value group is formed, and these groups are arranged in time order to obtain the response delay sequence.

[0101] In this embodiment of the invention, the voltage change rate sequence, current change rate sequence, and speed change rate sequence are divided into multiple window intervals according to the length of a preset time window, resulting in multiple rate windows. This improves the ability to capture local abrupt changes and avoids the occlusion effect on local details during global modeling. Within each rate window, the corresponding voltage change rate sequence and speed change rate sequence are extracted, sequence deviation matching is performed, and the minimum delay time of speed relative to voltage is calculated to obtain the voltage delay value. This accurately quantifies the hysteresis effect of voltage change on speed response and eliminates shape errors caused by different amplitudes or disturbance morphologies. Within the same rate window, the current change rate sequence is extracted, and the same sequence deviation matching is performed to calculate the minimum delay time of speed relative to current, obtaining the current delay value. This further establishes the dynamic response timing relationship between current input and speed response, revealing the hysteresis behavior of the load response in the electromagnetic link. Based on the voltage delay value and current delay value within each rate window, a delay value group is formed and arranged in chronological order to obtain the response delay sequence. This completes the process of abstracting the time response characteristics between input-output pairs into numerical indicators, which serves as the direct input basis for subsequent abrupt change identification and predictive control.

[0102] Specifically, based on the preset time window length, the voltage change rate sequence, current change rate sequence, and speed change rate sequence are divided into multiple window intervals according to time sequence, resulting in multiple rate windows, including:

[0103] First, a time window length parameter is preset, denoted as . The unit is milliseconds or the number of sampling points. The specific value can be set according to the system sampling frequency and response expectations. For example, in a high-frequency sampling scenario, it can be set to 100 sampling points. Based on this, according to the time axis in the original rate evolution dataset, the voltage change rate sequence, current change rate sequence, and speed change rate sequence are divided into equal-length segments along the time dimension. The segmentation method can adopt a fixed window segmentation method, that is, setting the starting point of the window to 1. The termination point is Then, with sliding step size Swipe forward to create the next window. to And so on, until the entire data sequence is covered.

[0104] Within each time window, the voltage change rate sequence, current change rate sequence, and speed change rate sequence within the current time range are extracted, denoted as . ,in Indicates the first There are [number] windows. The number of data points in the sequence within each window is [number]. ,equal The corresponding number of sampling points. After this step is completed, a set of rate windows with consistent time sequence, uniform sampling length and standardized data structure is obtained. Each rate window will serve as an independent analysis unit for subsequent sequence delay matching and feature recognition.

[0105] Specifically, based on the voltage and current delay values ​​within each rate window, delay sets are formed and arranged in chronological order to obtain a response delay sequence, which includes:

[0106] Within each constructed rate window, delay matching operations are performed between the voltage change rate sequence and the speed change rate sequence, as well as between the current change rate sequence and the speed change rate sequence. Specifically, the voltage (or current) sequence is kept stationary, and the corresponding speed change rate sequence slides along the time axis point by point within the window to construct a series of time-shifted speed sequences. These are compared one by one with the reference voltage (or current) sequence. Normalization is performed to ensure that the values ​​are all between [0,1]. The mean absolute error under each sliding sequence is then calculated. The number of sliding offset sampling points corresponding to the smallest error is selected and converted into a time value by combining the sampling time interval. This time value is used as the delay time of voltage relative to speed (i.e., voltage delay value) or the delay time of current relative to speed (i.e., current delay value) within the window.

[0107] Obtain the voltage delay value for each window. With current delay value Then, they are encapsulated into a set of delay pairs, denoted as... Repeat the above process, performing the matching calculation operation sequentially for all rate windows, and finally arrange the delay sets of all windows in time order to form a complete response delay sequence. This response delay sequence not only preserves the hysteresis characteristics of the input-output response within each time window, but also provides dynamic timing input features for subsequent operations such as identifying whether there are response abrupt changes, performing predictive modeling, and voltage compensation control.

[0108] In a preferred embodiment of the present invention, within each rate window, the corresponding voltage change rate sequence and speed change rate sequence are extracted, sequence deviation matching is performed, and the minimum delay time of speed relative to voltage is calculated to obtain the voltage delay value, including:

[0109] Based on the start and end times of the rate window, continuous data points of the voltage change rate sequence and the speed change rate sequence within the rate window are extracted to obtain voltage change rate window segments and speed change rate window segments.

[0110] Keeping the position of the voltage change rate window segment unchanged, slide the speed change rate window segment backward along the time axis within the current rate window by several sampling points to construct the first speed delay sequence;

[0111] The voltage change rate window segment and the first rotational speed delay sequence are normalized and their numerical range is scaled to the [0,1] interval to obtain the first normalized segment set;

[0112] Based on the first normalized fragment set, the average absolute error between the voltage and its corresponding rotational speed is calculated to obtain the voltage slip offset value data.

[0113] Based on the voltage sliding offset data, the number of sliding sampling points corresponding to the minimum voltage sliding offset value is identified. Combined with the sampling interval time, the offset time of the rotational speed relative to the voltage within the current rate window is determined, and the voltage delay value is obtained.

[0114] In this embodiment of the invention, based on the start and end times of the rate window, continuous data points of the voltage change rate sequence and the speed change rate sequence within the rate window are extracted to obtain voltage change rate window segments and speed change rate window segments. This ensures that voltage and speed are matched and analyzed within the same time interval, maintaining time consistency, which is the basis for subsequent time delay calculations. Keeping the position of the voltage change rate window segment unchanged, the speed change rate window segment is slid backward along the time axis within the current rate window by several sampling points to construct a first speed delay sequence. Multiple delayed versions of the speed segment are generated by this sliding motion, providing multiple candidate combinations for subsequent error comparison with the voltage segment. The voltage change rate window segment and the first speed delay sequence are normalized, scaling their numerical range to [specific values ​​would be inserted here]. In the [0,1] interval, a first normalized fragment set is obtained to eliminate the deviation caused by the difference in dimensions between different data sources and ensure that the voltage and speed fragments have the same comparison benchmark. Based on the first normalized fragment set, the average absolute error between the voltage and its corresponding speed is calculated to obtain the voltage sliding offset value data. The voltage-speed delay coupling relationship is quantified through error analysis to avoid the error interference of subjective experience judgment. Based on the voltage sliding offset value data, the number of sliding sampling points corresponding to the minimum voltage sliding offset value is identified. Combined with the sampling interval time, the offset time of speed relative to voltage within the current rate window is determined to obtain the voltage delay value. The minimum error matching result is transformed into a time delay value with physical meaning to realize the mapping from data difference to response characteristic index.

[0115] Specifically, keeping the position of the voltage change rate window segment unchanged, the rotational speed change rate window segment is slid backward along the time axis within the current rate window by several sampling points to construct the first rotational speed delay sequence, which includes:

[0116] First, the voltage change rate window segment has been extracted from the rate window. Its data consists of a set of equally spaced change rate values, denoted as... Meanwhile, the original speed window segment extracted from the speed change rate sequence is... Its initial position is aligned with the voltage segment. To simulate the delay effect of voltage change on the rotational speed response, the position of the voltage change rate window segment is not changed in this step; it remains stationary at the original rate window's starting point. Instead, the rotational speed window segment is slid forward point by point along the time axis, constructing multiple copies of the rotational speed sequence with different time delays. For example, the rotational speed segment is slid forward by 1, 2, 3, etc., from the original position. Each sampling point forms new sequence: Until These constitute the first set of rotational speed delay sequences, each sequence representing a rotational speed response pattern with a specific lag time (i.e., sliding steps × sampling interval).

[0117] Specifically, based on the first normalized fragment set, the average absolute error between the voltage and its corresponding rotational speed is calculated to obtain the voltage slip offset data, which includes:

[0118] Each speed delay segment and voltage change rate window segment is normalized using a linear scaling method. The minimum value of each segment is mapped to 0, the maximum value to 1, and the remaining points are calculated linearly to ensure they all fall within the [0,1] interval, thus eliminating the influence of the original data's dimensions. After normalization, let the normalized voltage segment be... Normalized The rotational speed delay segment is Next, the absolute error of each speed delay segment and voltage segment is calculated point by point, that is, the calculation... ,in Values ​​range from 0 to , Using the sliding step index, after completing this calculation, a voltage sliding offset value data array can be obtained, denoted as . It is used to determine the error level of voltage response to rotational speed under different delay states. The smaller the error value, the higher the degree of matching.

[0119] Specifically, based on the voltage slip offset data, the number of slip sampling points corresponding to the minimum voltage slip offset value is identified. Combined with the sampling interval time, the offset time of the rotational speed relative to the voltage within the current rate window is determined, resulting in the voltage delay value. This includes:

[0120] After obtaining the aforementioned voltage sliding offset data, a traversal search is performed to identify the minimum value and its corresponding index position. Let... To minimize the error value, This is the number of sliding sampling points corresponding to the optimal match between voltage and rotational speed. Subsequently, according to the preset system sampling interval (denoted as...), ), slide steps and By multiplying them, the time delay of the rotational speed response relative to the voltage change can be calculated. ,Right now This value, known as the voltage delay value, is a crucial indicator describing the dynamic relationship between voltage changes and speed response. This delay value is subsequently recorded in the response delay sequence and compared with a preset delay group to determine if any sudden changes in response occur. It can also serve as a timing benchmark for subsequent response modeling and offset prediction, ensuring the control system exhibits strong logical consistency and temporal stability.

[0121] In a preferred embodiment of the present invention, within the same rate window, the current change rate sequence is extracted, the same sequence deviation matching is performed, and the minimum delay time of rotational speed relative to current is calculated to obtain the current delay value, including:

[0122] Within the same rate window, extract continuous data points of the current change rate sequence within that rate window to obtain a current change rate window segment.

[0123] Keeping the position of the current rate of change window segment unchanged, the rotational speed rate of change window segment is slid backward along the time axis by several sampling points within the current rate window to construct a second rotational speed delay sequence;

[0124] The current change rate window segment and the second rotational speed delay sequence are normalized to scale their numerical range to the [0,1] interval, thus obtaining the second normalized segment set;

[0125] Based on the second normalized fragment set, the average absolute error between the current and its corresponding rotational speed is calculated to obtain the current slip offset value data.

[0126] Based on the current sliding offset value data, the number of sliding sampling points corresponding to the minimum current sliding offset value is identified. Combined with the sampling interval time, the offset time of the rotation speed relative to the current within the current rate window is determined, and the current delay value is obtained.

[0127] In this embodiment of the invention, within the same rate window, continuous data points of the current change rate sequence within that rate window are extracted to obtain a current change rate window segment. This dynamic signal of the current change rate is then subjected to time localization processing, allowing subsequent matching analysis to focus on local response characteristics and reducing the probability of misjudgment due to global interference or historical data inertia. Keeping the position of the current change rate window segment unchanged, the rotational speed change rate window segment is slid backward along the time axis within the current rate window by several sampling points to construct a second rotational speed delay sequence. This simulates the rotational speed response delay behavior under various hypothetical conditions, providing a basis for subsequent error matching. The current change rate window segment and the second rotational speed delay sequence are then normalized. The numerical range is scaled to the [0,1] interval to obtain a second normalized fragment set, ensuring that the comparison benchmark between current and speed is consistent and avoiding error judgment deviations due to different sampling amplitudes. Based on the second normalized fragment set, the average absolute error between the current and its corresponding speed is calculated to obtain current sliding offset value data, which effectively reflects the degree of offset between speed response and current change. Based on the current sliding offset value data, the number of sliding sampling points corresponding to the minimum current sliding offset value is identified, and combined with the sampling interval time, the offset time of speed relative to current within the current rate window is determined to obtain the current delay value, quantifying the minimum delay time of speed relative to current response, reflecting the actual response mechanism of the motor system.

[0128] Specifically, based on the current sliding offset value data, the number of sliding sampling points corresponding to the minimum current sliding offset value is identified. Combined with the sampling interval time, the offset time of the rotational speed relative to the current within the current rate window is determined, resulting in the current delay value. This includes:

[0129] First, the sliding error data is analyzed. The constructed sliding offset data is essentially a sequence of average absolute errors corresponding to different sliding sampling points. This sequence reflects the degree of matching between the speed change rate segment and the current change rate segment under different assumed delay conditions. The smaller the error, the more consistent the response trends of the two under that sliding condition. Therefore, based on this sliding offset data, the system can use the global minimum error criterion to locate the point with the smallest error among the sliding sampling points. This sliding point is the speed response position under the optimal alignment state, marking the actual response delay of the speed relative to the current.

[0130] Next, the number of sliding sampling points corresponding to the minimum offset value is recorded as follows: This parameter represents the number of sampling periods required for the rotational speed response to lag relative to the current change within the current window. Since the sliding process is constructed based on discrete sampling points, it still needs to be considered in conjunction with the sampling interval time set in the system. Time-scale mapping is performed on this offset number, i.e., through... The discrete sampled displacement is converted into a delay value in a continuous-time sense. This current delay value represents the minimum response time required for the motor speed to return to its normal range within a window after a change in the current signal. It is an empirical quantitative result of the motor's internal dynamic characteristics and response inertia.

[0131] In a preferred embodiment of the present invention, by calculating the difference between the response delay sequence and a preset delay group, it is determined whether there is a sudden change in the response of the brushed motor, and the corresponding abnormal response time is recorded to obtain a sudden change identification sequence, including:

[0132] Based on the voltage delay value and current delay value of each rate window in the response delay sequence, the difference between them and the preset voltage delay reference value and the preset current delay reference value is calculated to obtain the voltage delay difference and the current delay difference.

[0133] The absolute values ​​of the voltage delay difference and the current delay difference are processed to obtain non-negative delay offset pairs;

[0134] Determine whether any component in the non-negative delay offset value pair exceeds the preset offset threshold. If the result is yes, then there is a sudden change in response within the rate window, and the start time of the rate window is marked as the moment of abnormal response.

[0135] Arrange all abnormal response moments in chronological order to obtain the mutation identification sequence.

[0136] In this embodiment of the invention, the voltage delay value and current delay value of each rate window in the response delay sequence are compared with preset voltage delay reference values ​​and preset current delay reference values ​​to obtain voltage delay difference and current delay difference. By using the reference value difference calculation method, the offset between the actual response delay and the expected response state is clearly quantified, forming a continuous and comparable data sequence, which facilitates the subsequent unified standard for judging abnormal situations. The absolute value of the voltage delay difference and current delay difference is processed to obtain non-negative delay offset value pairs, removing positive and negative interference, which facilitates the setting of a unified mutation judgment threshold. It is judged whether any component in the non-negative delay offset value pair exceeds the preset offset threshold. When the result is yes, there is a response mutation within the rate window, and the start time of the rate window is marked as the response anomaly moment, realizing the discretization identification of the abnormal response window and effectively solving the false alarm problem under the influence of multi-parameter coupling. All response anomaly moments are arranged in chronological order to obtain a mutation identification sequence. The data path from the perception layer to the execution layer has a clear transmission structure and a unified format, providing time anchors for subsequent prediction modeling, voltage correction and compensation control modules.

[0137] The preset delay group includes a preset voltage delay reference value and a preset current delay reference value, which serve as reference standards for judging whether there is a sudden change in the response of the motor operating state. The two are set based on the actual operating characteristics of the brushed motor and a large amount of measured data under stable load conditions, combined with statistical analysis methods and engineering experience. They must ensure that there are no false alarms under normal operating conditions and have the ability to respond in a timely manner under abnormal conditions.

[0138] The preset voltage delay reference value reflects the minimum time offset required for the voltage change rate sequence to respond to the speed change rate sequence under normal load and stable operation. Specifically, multiple operating cycles under typical load conditions can be selected, and the delay value corresponding to the minimum sliding error position between the voltage change rate and the speed change rate in each cycle can be extracted. After multi-window and multi-sample statistics, the mean or mode is taken as the voltage response delay reference under steady-state conditions. For example, when an electric screwdriver is running under no-load or light-load conditions, voltage adjustment can usually cause a significant speed change within 1–2 sampling intervals. If the sampling period is 2ms, the reference value can be set between 2–4ms, that is, 1–2 time steps.

[0139] The preset current delay reference value reflects the time coupling delay between the electromagnetic response characteristics inside the motor and the current feedback of load changes. Under static load or constant torque, the current change is usually more rapid, so its reference value should be less than the voltage delay reference value. By sampling and analyzing the time difference between the current change rate and the speed response under a large number of normal operating conditions, the most common or average minimum delay time is obtained as the setting value. This value can be set to about 1–2 ms, and even in some high-speed response systems, it can be set to within 1 sampling point (i.e., 2 ms).

[0140] The values ​​of both should not be fixed to a certain constant, but should have parameterizable adjustment capabilities. In actual deployment, they can be automatically calibrated based on different motor models, control frequencies, load inertia and sensor response times through preset curves or system identification algorithms.

[0141] Specifically, the absolute values ​​of the voltage delay difference and the current delay difference are processed to obtain non-negative delay offset pairs, including:

[0142] For each time window, the absolute values ​​of the voltage delay difference and current delay difference corresponding to that window are transformed. The mathematical expression is as follows: For each window The voltage delay difference Transform into Current delay difference Transform into These two values ​​represent the absolute degree of voltage response delay and current response delay offset, respectively, without considering the direction of the offset. Finally, these two non-negative values ​​are combined into an ordered pair, forming the non-negative delay offset value pair corresponding to this window.

[0143] In a preferred embodiment of the present invention, based on the mutation identification sequence, the rate evolution data corresponding to each abnormal response moment is extracted, and the weighted average rotational speed change rate is calculated to obtain the predicted rotational speed offset sequence, including:

[0144] Based on the abnormal response moments in the mutation identification sequence, the rotational speed change rate, current change rate, and voltage change rate data within a fixed prediction time starting from any abnormal response moment are extracted from the rate evolution dataset to obtain the abnormal response dataset.

[0145] Based on the response anomaly dataset, the weighting factors of the current change rate and voltage change rate are calculated, and combined with the speed change rate data, the weighted average speed change rate at the moment of the response anomaly is calculated.

[0146] The predicted speed offset is obtained based on the weighted average speed change rate and a fixed prediction duration.

[0147] Arrange the predicted speed offsets corresponding to all abnormal response times in chronological order to obtain the predicted speed offset sequence.

[0148] In this embodiment of the invention, based on the abnormal response moments in the mutation identification sequence, speed change rate data, current change rate data, and voltage change rate data within a fixed prediction duration starting from any abnormal response moment are extracted from the rate evolution dataset to obtain a response anomaly dataset. This dataset precisely focuses on the local dynamic process during sudden changes in motor state, ensuring consistency in the temporal and physical correlation of the prediction input data, thus laying a high-fidelity data foundation for subsequent calculations of speed change trends. Based on the response anomaly dataset, weighting factors for current and voltage change rates are calculated, and combined with the speed change rate data, a weighted average speed change rate at the abnormal response moment is calculated, reflecting the instantaneous changes in voltage and current. The method accurately measures the actual force intensity of time-varying motor responses, solving the problem of distorted speed responses caused by single-parameter drive in traditional methods. Based on the weighted average speed change rate and a fixed prediction duration, the predicted speed offset is obtained, ensuring the predicted values ​​have a reasonable source, data-driven logic, and time-domain validity, effectively supporting the quantitative basis required for subsequent compensation calculations. The predicted speed offsets corresponding to all abnormal response moments are arranged in chronological order to obtain a predicted speed offset sequence, realizing the transformation from discrete prediction points to continuous compensation control trajectories. This ensures the control system has responsive predictive references at every critical disturbance node and also provides quantitative support for smooth control strategy design.

[0149] In scenarios where load disturbances primarily manifest as abrupt changes in output-side mechanical impedance (e.g., the instant an electric screwdriver drives into a steel plate), the current signal, as the primary characteristic parameter of the load response, exhibits the strongest abrupt change, reflecting the sudden increase in the back electromotive force required by the rotor magnetic field. Therefore, in such high mechanical inertia conditions, the current weighting coefficient should be increased. The proportion of The values ​​are 0.55, 0.25 and 0.2 respectively, to ensure that the prediction model is more sensitive to the coupling relationship between current and speed, and is especially suitable for medium and high speed impact loads.

[0150] In applications where voltage disturbances significantly dominate response characteristics (such as small electric drills with frequent PWM duty cycle adjustments or during motor control strategy switching), drive-end voltage fluctuations often determine the rate of energy input change. In such cases, greater attention should be paid to voltage-speed coupling characteristics. The values ​​of are 0.3, 0.5 and 0.2 respectively. At this time, the model is more sensitive to the fluctuation of the input signal, and the prediction results focus more on the modeling performance of the power supply response speed, which is suitable for high-frequency control or soft start-stop scenarios.

[0151] In some operating conditions where the response to both supply and demand is balanced but the magnitude of changes is not drastic, such as in low-power everyday power tools or systems with partial speed stabilization capabilities, a relatively neutral and balanced weight allocation can be adopted. The values ​​of 0.4, 0.4 and 0.2 are respectively, which can maintain the stability of prediction modeling when the changes in system parameters do not show the dominant variable, reduce the abnormal prediction offset caused by extreme weight error amplification, and make the control logic more robust.

[0152] Regardless of the scenario The proportion of the constant adjustment term is generally maintained in the range of 0.15-0.25. Its function is to provide the minimum reference response for situations where the sudden change signal is not significant or the fluctuation of the two signals is close to zero, so as to avoid the denominator approaching 0 during the weighting process, thereby ensuring the numerical stability and mathematical completeness of the system. The final weight value should be optimized based on the historical operating condition data of the application equipment through experimental sampling or online learning.

[0153] In a preferred embodiment of the present invention, based on the response anomaly dataset, weighting factors for the current change rate and voltage change rate are calculated, and combined with the speed change rate data, a weighted average speed change rate at the moment of the response anomaly is calculated, including:

[0154] Within a fixed prediction time at any time of response anomaly, the rate of change of current at each sampling point in the response anomaly segment is read sequentially, and the current weight value is obtained based on its absolute value and the preset current adjustment value.

[0155] The voltage change rate of each sampling point in the same response anomaly segment is read sequentially, and the voltage weight value is obtained based on its absolute value and the preset voltage adjustment value.

[0156] The current weight value, voltage weight value, and preset constant adjustment value corresponding to each sampling point are fused together to obtain a comprehensive weighting factor;

[0157] The weighted rate of change of rotational speed is obtained by multiplying the rate of change of rotational speed at each sampling point with its corresponding comprehensive weighting factor.

[0158] The weighted average speed change rate is obtained by summing all weighted speed change rates within the abnormal response segment and dividing it by the sum of all comprehensive weighting factors within the abnormal response segment.

[0159] In this embodiment of the invention, within a fixed prediction time at any abnormal response moment, the current change rate of each sampling point in the abnormal response segment is sequentially read, and a current weight value is obtained based on its absolute value and a preset current adjustment value, effectively identifying the dominant influence of current change on rotational speed during abrupt changes; the voltage change rate of each sampling point in the same abnormal response segment is sequentially read, and a voltage weight value is obtained based on its absolute value and a preset voltage adjustment value, ensuring that the system fully considers the excitation effect of the power supply side in abnormal response judgment; the current weight value, voltage weight value, and preset constant adjustment value corresponding to each sampling point are fused to obtain a comprehensive weighting factor, realizing the... The organic unification of the influence of multiple source variables overcomes the problem that a single weight may lead to amplified bias. Multiplying the rotational speed change rate of each sampling point with its corresponding comprehensive weighting factor yields the weighted rotational speed change rate, so that each data point is no longer treated equally, but is differentiated based on response capability, thus suppressing the system's overfitting to irrelevant changes. Summing all weighted rotational speed change rates within the abnormal response segment and dividing it by the sum of all comprehensive weighting factors within the abnormal response segment yields the weighted average rotational speed change rate. While filtering out the influence of random disturbances, it retains the key components that truly affect the evolution of the system's rotational speed, avoiding the risk of single-point values ​​dominating the prediction.

[0160] The preset current adjustment value, preset voltage adjustment value, and preset constant adjustment value correspond to the adjustment and control of the current change rate, voltage change rate, and system basic balance factor, respectively, and are important weight benchmarks for constructing the influence distribution of multi-parameter response.

[0161] Preset current adjustment value This function is primarily used to normalize the absolute value of the current change rate. Its purpose is to establish a standard scale for comparison of the current response across different sampling points. By comparing the current change rate at a given moment with this preset value, the current weight is calculated, allowing the system to determine whether the current fluctuation at that sampling point is significant. If the current response is too drastic and the preset current adjustment value is too small, the weight will be too high, amplifying the current's influence in the model. Conversely, if the preset current adjustment value is set too large, the contribution of current change to the weighted speed change rate will be weakened. Therefore, the specific value of the preset current adjustment value should be set in conjunction with the motor current fluctuation range under typical loads. For example, in a conventional electric screwdriver, the current change range per unit time is typically ±1.5A. The preset current adjustment value can be set to 1A, ensuring coverage of normal operating fluctuations while providing good resolution for sudden changes.

[0162] Preset voltage adjustment value In calculating voltage weights, it serves an equivalent function to the preset current regulation value. It standardizes the absolute value of the voltage change rate to assess the degree of dominance exhibited by the voltage application end during abrupt changes in response. The preset voltage regulation value, as a normalization factor, should reflect the effective response range of the average voltage fluctuation in motor speed control. In PWM speed control systems, the common voltage change rate is generally around ±2V / ms; therefore, the preset voltage regulation value can be selected to be around 1.5V / ms to ensure the objectivity and stability of this parameter in judging voltage dominance within the abrupt change identification window.

[0163] Preset constant adjustment value Its main function is to introduce a reference coefficient that still provides basic control weights when voltage and current fluctuations are not significant, thus balancing the model's output stability in scenarios with weak signal fluctuations. This constant adjustment value can be set to a constant value between the voltage and current normalization weights, with a recommended initial value of 1 to facilitate forming a reference value of the same order of magnitude as other normalization ratios. In actual control, it can also be dynamically adjusted to appropriately correct for sudden changes in frequency, thereby avoiding prediction bias caused by over-reliance on voltage or current variations.

[0164] In summary, the functions of these three preset values ​​are to provide normalized references for current and voltage change rates, and to provide stable benchmarks for the construction of weighting factors. Their values ​​are mainly based on statistical analysis of the range of measured motor operation data, power supply regulation mechanisms, and algorithm tolerance design. Experimental calibration and parameter calibration are required before system integration to adapt to the actual operating characteristics of specific motor models and load types.

[0165] In a preferred embodiment of the present invention, the compensation voltage increment is calculated based on the difference between the predicted speed offset sequence and the real-time speed to obtain corrected voltage data, including:

[0166] Based on the predicted speed offset sequence, extract the predicted speed offset at any time of response anomaly, and extract the real-time speed at the current time.

[0167] The difference between the predicted speed offset and the real-time speed is calculated to obtain the speed compensation requirement value.

[0168] Within a fixed prediction time at the moment of the response anomaly, the voltage change rate and the speed change rate are extracted, and the average corresponding ratio between the two is calculated to obtain the response conversion factor.

[0169] Calculate the compensation voltage increment based on the response conversion factor and the speed compensation requirement;

[0170] The corrected voltage data is obtained based on the compensation voltage increment and the control voltage at the current moment.

[0171] In this embodiment of the invention, based on the predicted speed offset sequence, the predicted speed offset at any abnormal response moment is extracted, and the real-time speed at the current moment is also extracted. This helps to form a quantitative description of the system state offset under the same time reference, providing a basis for judgment for subsequent error-driven compensation mechanisms. The difference between the predicted speed offset and the real-time speed is calculated to obtain the speed compensation requirement value, effectively avoiding the risk of misadjustment caused by local fluctuations. Within a fixed prediction time of the abnormal response moment, the voltage change rate and speed change rate are extracted, and the average corresponding ratio between the two is calculated to obtain the response conversion factor, constructing a nonlinear coefficient that maps speed error to voltage increment. Based on the response conversion factor and the speed compensation requirement value, the compensation voltage increment is calculated, and the dynamic conversion of speed offset to voltage compensation is realized through parameter coupling, forming a direct path from the predicted trend to the voltage correction amount. Based on the compensation voltage increment and the control voltage at the current moment, the correction voltage data is obtained. This correction value realizes the organic combination of steady-state adjustment and sudden intervention, effectively maintaining the stable operating state of the brushed motor speed.

[0172] In operating conditions applicable to most power tools with frequent starts and stops and load variations that are not extreme but fluctuate, current, as the most sensitive indicator of load changes, usually exhibits peak fluctuations first at the moment of abrupt change at the mechanical end, and therefore should be given the highest weight; voltage, as the main means by which the control system applies drive signals, directly reflects the adjustment intention, but its response speed is slightly slower than that of current, so its weight is moderate; constant items... The introduction of this feature aims to retain a certain response basis when voltage and current signals tend to be stable or close to zero rate of change, ensuring that even when the system is in a perturbation or quasi-steady state, it can still generate a moderately compensated signal. The values ​​are 0.5, 0.3, and 0.2.

[0173] When the controlled object is in a typical impact scenario with drastic load changes, such as an electric impact drill or hammer drill, the system's sensitivity to current response should be significantly enhanced, enabling faster judgment and adjustment under high dynamic characteristics. Simultaneously, the weight of the voltage change rate should be appropriately reduced to avoid overly aggressive adjustments that cause voltage fluctuations, while the constant bias should be slightly reduced to improve the resolution of the judgment threshold, making the compensation process more compact and the response more decisive. The values ​​are 0.6, 0.25, and 0.15.

[0174] In scenarios requiring high steady-state accuracy but with gradually changing loads, such as in micro electric actuators and precision servo control systems, fine-tuning is relied upon to maintain a constant speed, and current fluctuations are not drastic, thus reducing its dominant role. Regulation is primarily achieved through high-frequency voltage PWM fine-tuning, thus giving voltage a higher weight. Constant terms become key factors in this scenario, enhancing the system's regulatory sensitivity during perturbations, thereby preventing quiescent states and error accumulation. The values ​​are 0.3, 0.4, and 0.3 respectively.

[0175] In a preferred embodiment of the present invention, a compensation voltage signal is generated based on the corrected voltage data and transmitted to the drive input terminal of the brushed motor to ensure stable motor speed, including:

[0176] Calculate the change in the corrected voltage data compared to the control voltage at the previous moment to obtain the correction magnitude;

[0177] If the correction magnitude exceeds the preset voltage adjustment threshold, the correction magnitude will be limited to ensure that the correction magnitude does not exceed the preset voltage adjustment threshold.

[0178] Determine whether the corrected voltage data exceeds the brushed motor voltage output range. If the result is yes, then limit the corrected voltage data to the brushed motor voltage output range.

[0179] After the above restrictions, the restricted correction voltage data is obtained. The weighted average of the restricted correction voltage data and the corrected voltage data is then used to obtain the smoothed correction voltage data.

[0180] Based on the smoothed correction voltage data, its value is converted into a voltage output command, a compensation voltage signal is generated, and it is transmitted to the drive input terminal of the brushed motor.

[0181] In this embodiment of the invention, the change in the corrected voltage data compared to the control voltage at the previous moment is calculated to obtain the correction amplitude. This quantifies the output adjustment amplitude of the controller within the current control cycle, facilitating the identification of any voltage surges. When the correction amplitude exceeds a preset voltage adjustment threshold, it is limited to ensure that the correction amplitude does not exceed the preset voltage adjustment threshold. This effectively suppresses the impact of sudden large voltage changes on the motor windings or drive circuit, while avoiding motor overshoot caused by overly aggressive control. It is also determined whether the corrected voltage data exceeds the brushed motor voltage output range. If the result is yes, the corrected voltage data is limited to the brushed motor voltage output range. To ensure that the control output is within the electrical range allowed by the physical equipment, and to prevent the controller output from exceeding the limits due to inaccurate prediction or regulation, the above-mentioned limiting correction voltage data is obtained. This data is then weighted and averaged with the correction voltage data to obtain smoothed correction voltage data. This smoothed correction voltage data can eliminate voltage fluctuations caused by disturbance estimation bias or transient prediction error, effectively avoiding sudden changes in the voltage output signal. Based on the smoothed correction voltage data, its value is converted into a voltage output command, generating a compensation voltage signal, which is then transmitted to the drive input of the brushed motor. This ensures the consistency of response between data calculation and actual physical action, effectively realizing active intervention and dynamic voltage regulation of speed fluctuations.

[0182] Specifically, when the correction magnitude exceeds a preset voltage adjustment threshold, the correction magnitude is limited to ensure that it does not exceed the preset voltage adjustment threshold. This includes:

[0183] Within each control cycle, the corrected voltage data at the current moment and the control voltage data at the previous moment are acquired, and the difference between the two is calculated to obtain the correction amplitude. Subsequently, a preset voltage adjustment threshold is set to characterize the maximum allowable adjustment range of the voltage per unit time. If the absolute value of the correction amplitude is greater than the preset threshold, it indicates that the current voltage change is too fast or too large, which may cause overshoot in the motor system response or abnormal output of the driver. Therefore, the correction amplitude is limited, that is, the correction amplitude is adjusted to a correction value whose sign remains unchanged and whose absolute value is equal to the preset voltage adjustment threshold, thereby obtaining the limited voltage correction result and ensuring that the correction amplitude is within the safe adjustment range of the system. If the absolute value of the correction amplitude does not exceed the preset voltage adjustment threshold, the original corrected voltage data is kept unchanged, thereby completing the dynamic control and suppression of rapid voltage fluctuations.

[0184] Specifically, it involves determining whether the corrected voltage data exceeds the brushed motor's voltage output range. If the result is yes, the corrected voltage data is then limited to the brushed motor's voltage output range. This includes:

[0185] The voltage value obtained after the aforementioned limiting process Next, based on the rated operating range of the brushed motor used, the minimum acceptable operating voltage for the motor is set. With maximum operating voltage ,in Used to limit motor starting failure or insufficient output torque caused by excessively low voltage. This is used to prevent the risk of motor insulation damage, winding overheating, or driver breakdown due to excessive voltage; in actual operation, it is compared and and If the relationship, Less than Then Adjusted to ;like Greater than Then Adjusted to ;like If the voltage is within this range, it remains unchanged, ultimately yielding the voltage correction data after output limiting. This ensures that the voltage value is always within the physical range allowed for normal operation of the brushed motor.

[0186] The above-mentioned constraints result in constrained corrected voltage data. This constrained corrected voltage data is then weighted and averaged with the corrected voltage data to obtain smoothed corrected voltage data. Specifically, this includes:

[0187] Corrected voltage after obtaining the above range limitations and the initial correction voltage Subsequently, a weighted averaging strategy is used to generate the final smoothed correction voltage used to control the output. Specifically, a smoothing factor is set. , The value ranges from 0 to 1, and is typically chosen between 0.7 and 0.95 to achieve a trade-off between response sensitivity and control stability; and By performing weighted fusion, the expression for calculating the smoothed correction voltage is obtained: If the current restrictions are relatively small, and If they are close, then and A small difference in voltage level maintains high system responsiveness; if the limiting range is large, smooth output will effectively suppress input jumps, achieve continuous voltage transition, and improve the drive stability of the motor under dynamic operating conditions.

[0188] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for controlling the speed of a brushed motor, characterized in that, The method includes: The voltage, current, and speed of the brushed motor were collected and time-aligned to obtain a ternary synchronization dataset. Based on the three-dimensional synchronization dataset, the rate of change of voltage, current and rotation speed at adjacent time points is calculated, and the corresponding rate of change sequences are constructed to obtain the rate evolution dataset. Based on a preset time window, the rate of change sequence in the rate evolution dataset is divided into multiple window intervals, and within each window interval, the delay quantity group is calculated to obtain the response delay sequence. By calculating the difference between the response delay sequence and the preset delay group, it is determined whether there is a sudden change in the response of the brushed motor, and the corresponding abnormal response time is recorded to obtain the change identification sequence. Based on the mutation identification sequence, the rate evolution data corresponding to each abnormal response moment is extracted, the weighted average speed change rate is calculated, and the predicted speed offset sequence is obtained. The compensation voltage increment is calculated based on the difference between the predicted speed offset sequence and the real-time speed to obtain the corrected voltage data; Based on the corrected voltage data, a compensation voltage signal is generated and transmitted to the drive input of the brushed motor to ensure stable motor speed. The formula for calculating the predicted rotational speed offset is as follows: ， in, For the first Predicted rotational speed offset at each abnormal response moment Index in response to abnormal events, To fix the prediction duration, This represents the total number of sampling points within a fixed prediction period. The index of the sampling points within a fixed prediction time period, For the first The rate of change of current at each sampling point For the first The rate of change of current at each sampling point For the first After the first abnormal response moment The rate of change of rotational speed at each sampling point To preset the current adjustment value, The preset voltage adjustment value, To preset constant adjustment value, These are the weighting coefficients, and their sum is 1; in, For the first The weighted average rate of change of rotational speed at each sampling point and The first Current weight value and voltage weight value at each sampling point For the first The comprehensive weighting factor for each sampling point For the first Weighted rate of change of rotational speed at each sampling point; The formula for calculating the compensation voltage increment is as follows: ， in, For the first The compensation voltage increment at each abnormal response moment For the first Predicted rotational speed offset at each abnormal response moment For the first Real-time rotation speed at each abnormal response moment To fix the number of sampling points within the prediction time period, The index of the sampling points within a fixed prediction time period, For the first The rate of change of current at each sampling point For the first Voltage change rate at each sampling point For the first The rate of change of rotational speed at each sampling point To preset the current adjustment value, The preset voltage adjustment value, This is a local minimum constant used to avoid division by zero errors; it is set to... , These are the weighting coefficients, and their sum is 1; The expression for calculating the smoothing correction voltage is as follows: ,in, To smooth the correction voltage, The corrected voltage after limitation For the initial correction voltage, It is a smoothing factor, and its value ranges from 0 to 1.

2. The method for controlling the speed of a brushed motor according to claim 1, characterized in that, Based on the ternary synchronous dataset, the rates of change of voltage, current, and rotational speed at adjacent time points are calculated, and corresponding rate of change sequences are constructed to obtain a rate evolution dataset, including: Based on the difference between the current voltage value and the previous voltage value in the three-dimensional synchronous dataset, and combined with the sampling interval time, the voltage change rate is calculated to obtain the voltage change rate sequence. Based on the difference between the current value at the current moment and the current value at the previous moment in the three-dimensional synchronous dataset, and combined with the sampling interval time, the current change rate is calculated to obtain the current change rate sequence. Based on the difference between the current rotational speed value and the previous rotational speed value in the three-dimensional synchronous dataset, and combined with the sampling interval time, the rotational speed change rate is calculated to obtain the rotational speed change rate sequence. A rate evolution dataset is constructed based on the voltage change rate sequence, current change rate sequence, and rotational speed change rate sequence.

3. The method for controlling the speed of a brushed motor according to claim 2, characterized in that, Based on the time window, the rate of change sequence in the rate evolution dataset is divided into multiple window intervals, and within each window interval, the delay quantity group is calculated to obtain the response delay sequence, including: Based on the preset time window length, the voltage change rate sequence, current change rate sequence, and speed change rate sequence are divided into multiple window intervals according to time sequence to obtain multiple rate windows; Within each rate window, the corresponding voltage change rate sequence and speed change rate sequence are extracted, sequence deviation matching is performed, the minimum delay time of speed relative to voltage is calculated, and the voltage delay value is obtained. Within the same rate window, the current change rate sequence is extracted, the same sequence deviation matching is performed, the minimum delay time of rotational speed relative to current is calculated, and the current delay value is obtained. Based on the voltage delay value and current delay value within each rate window, a delay value group is formed, and these groups are arranged in time order to obtain the response delay sequence.

4. The method for controlling the speed of a brushed motor according to claim 3, characterized in that, Within each rate window, the corresponding voltage change rate sequence and speed change rate sequence are extracted, sequence deviation matching is performed, and the minimum delay time of speed relative to voltage is calculated to obtain the voltage delay value, including: Based on the start and end times of the rate window, continuous data points of the voltage change rate sequence and the speed change rate sequence within the rate window are extracted to obtain voltage change rate window segments and speed change rate window segments. Keeping the position of the voltage change rate window segment unchanged, slide the speed change rate window segment backward along the time axis within the current rate window by several sampling points to construct the first speed delay sequence; The voltage change rate window segment and the first rotational speed delay sequence are normalized and their numerical range is scaled to the [0,1] interval to obtain the first normalized segment set; Based on the first normalized fragment set, the average absolute error between the voltage and its corresponding rotational speed is calculated to obtain the voltage slip offset value data. Based on the voltage sliding offset data, the number of sliding sampling points corresponding to the minimum voltage sliding offset value is identified. Combined with the sampling interval time, the offset time of the rotational speed relative to the voltage within the current rate window is determined, and the voltage delay value is obtained.

5. The method for controlling the speed of a brushed motor according to claim 4, characterized in that, Within the same rate window, the current change rate sequence is extracted, and the same sequence deviation matching is performed. The minimum delay time of rotational speed relative to current is calculated to obtain the current delay value, including: Within the same rate window, extract continuous data points of the current change rate sequence within that rate window to obtain a current change rate window segment. Keeping the position of the current rate of change window segment unchanged, the rotational speed rate of change window segment is slid backward along the time axis by several sampling points within the current rate window to construct a second rotational speed delay sequence; The current change rate window segment and the second rotational speed delay sequence are normalized to scale their numerical range to the [0,1] interval, thus obtaining the second normalized segment set; Based on the second normalized fragment set, the average absolute error between the current and its corresponding rotational speed is calculated to obtain the current slip offset value data. Based on the current sliding offset value data, the number of sliding sampling points corresponding to the minimum current sliding offset value is identified. Combined with the sampling interval time, the offset time of the rotation speed relative to the current within the current rate window is determined, and the current delay value is obtained.

6. The method for controlling the speed of a brushed motor according to claim 5, characterized in that, By calculating the difference between the response delay sequence and the preset delay group, it is determined whether there is a sudden change in the response of the brushed motor, and the corresponding abnormal response time is recorded to obtain the change identification sequence, including: Based on the voltage delay value and current delay value of each rate window in the response delay sequence, the difference between them and the preset voltage delay reference value and the preset current delay reference value is calculated to obtain the voltage delay difference and the current delay difference. The absolute values ​​of the voltage delay difference and the current delay difference are processed to obtain non-negative delay offset pairs; Determine whether any component in the non-negative delay offset value pair exceeds the preset offset threshold. If the result is yes, then there is a sudden change in response within the rate window, and the start time of the rate window is marked as the moment of abnormal response. Arrange all abnormal response moments in chronological order to obtain the mutation identification sequence.

7. The method for controlling the speed of a brushed motor according to claim 6, characterized in that, Based on the mutation identification sequence, the rate evolution data corresponding to each abnormal response moment is extracted, and the weighted average speed change rate is calculated to obtain the predicted speed offset sequence, including: Based on the abnormal response moments in the mutation identification sequence, the rotational speed change rate, current change rate, and voltage change rate data within a fixed prediction time starting from any abnormal response moment are extracted from the rate evolution dataset to obtain the abnormal response dataset. Based on the response anomaly dataset, the weighting factors of the current change rate and voltage change rate are calculated, and combined with the speed change rate data, the weighted average speed change rate at the moment of the response anomaly is calculated. The predicted speed offset is obtained based on the weighted average speed change rate and a fixed prediction duration. Arrange the predicted speed offsets corresponding to all abnormal response times in chronological order to obtain the predicted speed offset sequence.

8. The method for controlling the speed of a brushed motor according to claim 7, characterized in that, Based on the response anomaly dataset, weighting factors for the current and voltage change rates are calculated, and combined with the speed change rate data, the weighted average speed change rate at the time of the response anomaly is calculated, including: Within a fixed prediction time at any time of response anomaly, the rate of change of current at each sampling point in the response anomaly segment is read sequentially, and the current weight value is obtained based on its absolute value and the preset current adjustment value. The voltage change rate of each sampling point in the same response anomaly segment is read sequentially, and the voltage weight value is obtained based on its absolute value and the preset voltage adjustment value. The current weight value, voltage weight value, and preset constant adjustment value corresponding to each sampling point are fused together to obtain a comprehensive weighting factor; The weighted rate of change of rotational speed is obtained by multiplying the rate of change of rotational speed at each sampling point with its corresponding comprehensive weighting factor. The weighted average speed change rate is obtained by summing all weighted speed change rates within the abnormal response segment and dividing it by the sum of all comprehensive weighting factors within the abnormal response segment.

9. The method for controlling the speed of a brushed motor according to claim 8, characterized in that, Based on the difference between the predicted speed offset sequence and the real-time speed, the compensation voltage increment is calculated to obtain the corrected voltage data, including: Based on the predicted speed offset sequence, extract the predicted speed offset at any time of response anomaly, and extract the real-time speed at the current time. The difference between the predicted speed offset and the real-time speed is calculated to obtain the speed compensation requirement value. Within a fixed prediction time at the moment of the response anomaly, the voltage change rate and the speed change rate are extracted, and the average corresponding ratio between the two is calculated to obtain the response conversion factor. Calculate the compensation voltage increment based on the response conversion factor and the speed compensation requirement; The corrected voltage data is obtained based on the compensation voltage increment and the control voltage at the current moment.

10. The method for controlling the speed of a brushed motor according to claim 9, characterized in that, Based on the corrected voltage data, a compensation voltage signal is generated and transmitted to the drive input of the brushed motor to ensure stable motor speed, including: Calculate the change in the corrected voltage data compared to the control voltage at the previous moment to obtain the correction magnitude; If the correction magnitude exceeds the preset voltage adjustment threshold, the correction magnitude will be limited to ensure that the correction magnitude does not exceed the preset voltage adjustment threshold. Determine whether the corrected voltage data exceeds the brushed motor voltage output range. If the result is yes, then limit the corrected voltage data to the brushed motor voltage output range. After the above restrictions, the restricted correction voltage data is obtained. The weighted average of the restricted correction voltage data and the corrected voltage data is then used to obtain the smoothed correction voltage data. Based on the smoothed correction voltage data, its value is converted into a voltage output command, a compensation voltage signal is generated, and it is transmitted to the drive input terminal of the brushed motor.

Images