An adaptive control method and system for a multi-channel electro-particle transducer

Abstract

This invention discloses an adaptive control method and system for a multi-channel electroporator, relating to the field of cell engineering instrument control technology. The method involves selecting a diagnostic model based on the cell suspension type; acquiring electrical data from each channel in real time and extracting feature vectors; using the diagnostic model for detection to obtain the detection results for each channel; comprehensively judging the operating status of the electroporator based on the results of all channels; and implementing a correction or shutdown strategy for abnormal channels according to different statuses. Through online analysis and real-time diagnosis, this invention can instantly identify and distinguish abnormal channels during the experiment, thereby adaptively compensating for correctable channels or promptly shutting down faulty channels, effectively avoiding waste of samples and reagents, and significantly improving the reliability and intelligence level of multi-channel electroporation experiments.

Description

Technical Field

[0001] This invention relates to the field of cell engineering instrument control technology, specifically to an adaptive control method and system for a multichannel electroporator. Background Technology

[0002] Multichannel electroporators, as high-throughput cell processing devices, are widely used in gene transfection, cell engineering, and related biological experiments. Their design goal is to improve experimental efficiency and reduce operating costs through the parallel operation of multiple working channels. Under current technological conditions, multichannel electroporation processes are typically completed based on preset electroporation parameters. The equipment mainly executes a predetermined process around control quantities such as voltage amplitude and pulse timing. The judgment of the operating status of each channel during electroporation largely depends on the transfection results or endpoint detection data after the experiment is completed.

[0003] In practical applications, cell electroporation exhibits significant procedural characteristics, with changes in the internal state of the channels directly impacting cell viability and transfection efficacy. However, current technologies primarily use process data for recording or basic monitoring, focusing on safety issues such as short circuits and overheating, while neglecting situations where researchers mismatch the electroporation program with the cell sample. No diagnostic conclusions are generated for the operational status of individual channels during the experiment; researchers typically need to complete electroporation of all channels and perform subsequent testing before determining whether the sample yielded valid results. This outcome-oriented evaluation method makes it impossible to identify or address anomalies that occur during the experiment in the current batch, leading to sample waste. Summary of the Invention

[0004] The purpose of this invention is to solve the problem mentioned in the background art that the existing technology cannot diagnose the channel status in real time during the electroporation process, resulting in abnormal channels not being distinguished and dealt with in a timely manner, and causing sample waste. Therefore, this invention proposes an adaptive control method and system for multi-channel electroporators.

[0005] A first aspect of this invention provides an adaptive control method for a multi-channel electro-particle transducer, the method comprising: Obtain descriptive data of the cell suspension to be processed, and determine the corresponding diagnostic model based on the descriptive data; Real-time acquisition of electrical data for each working channel of the electro-particle converter; Feature vectors are obtained by extracting features from the electrical data of the target channel; the target channel is any working channel of the electro-particle converter. The feature vector is input into the diagnostic model to obtain the detection result of the target channel; The working status is determined by comprehensively judging the test results from multiple channels; the working status includes normal, single point abnormality, and batch abnormality. Based on the operating status, execute the corresponding processing strategy; the processing strategy includes correcting abnormal channels and closing abnormal channels.

[0006] By implementing this technical solution, real-time online diagnosis and adaptive control of channel status during electroporation are achieved. Abnormal channels can be identified and handled in a timely manner, effectively preventing sample waste caused by program mismatch or equipment failure, and improving experimental efficiency and reliability.

[0007] Optionally, one diagnostic model corresponds to one type of cell suspension; the construction process of each diagnostic model includes: Acquire electroporation feedback data to identify normal clusters in the target cell type suspension; the electroporation feedback data is input by the user after the experiment; the normal clusters include data records with multiple electroporation results showing normality; For each normal data record, feature extraction is performed to obtain a sample feature vector; Based on the sample feature vectors of multiple data records, a single-class support vector machine is trained to obtain the diagnostic model corresponding to the target type of cell suspension.

[0008] By implementing this technical solution, a normal model is constructed using the electrical characteristics of historical successful experiments, enabling sensitive detection of abnormal states that deviate from the normal pattern. This eliminates the need to rely on abnormal sample data that is difficult to obtain in advance, thus improving the feasibility of model construction and the accuracy of anomaly identification.

[0009] Optionally, executing the corresponding processing strategy based on the working state includes: If the working status is a single point of failure, then compensation is determined for each abnormal channel; if it can be compensated, then compensation is performed; otherwise, the abnormal channel that cannot be compensated is closed. If the working status is batch abnormal, then a consistency check is performed on all abnormal channels; if they are consistent, a warning message is displayed and a decision is made on whether to continue the electronic transfer operation based on the feedback; otherwise, all channels are shut down.

[0010] By implementing this technical solution, differentiated handling strategies were developed for different types of abnormal states. This approach can not only dynamically compensate for locally repairable anomalies to preserve samples, but also provide timely warnings or halt operations when systemic risks occur, thus achieving intelligent hierarchical processing.

[0011] Optionally, compensation determination and compensation for each abnormal channel include: Based on the normal channels of the current batch, calculate the average value of its current waveform to obtain the reference waveform; A similarity calculation is performed between the current waveform of the target abnormal channel and the reference waveform to obtain the waveform similarity; the target abnormal channel is any one of the channels identified as abnormal. If the waveform similarity is greater than a preset stability threshold, then the target abnormal channel is determined to be compensable. The compensation coefficient is calculated based on the average steady-state current of the normal channel and the steady-state current of the target abnormal channel, and the compensation coefficient is used to compensate the pulse voltage. Starting from the next pulse, a compensated pulse is applied to the target abnormal channel.

[0012] By implementing this technical solution, similarity analysis is performed between the current waveform of the abnormal channel and the normal reference waveform to intelligently distinguish between recoverable temporary deviations and irrecoverable severe anomalies. Real-time voltage adjustment is performed on the compensable channels, which maximizes the use of all available channels and reduces sample loss while ensuring the overall transfection effect.

[0013] Optionally, consistency checks on all abnormal channels include: Calculate the similarity of feature vectors between any two abnormal channels to obtain multiple sets of similarity data; Calculate the average value of the multiple sets of similarity data to obtain the overall similarity. If the overall similarity is greater than the preset consistency threshold, all abnormal channels are determined to have consistent features.

[0014] By implementing this technical solution and comprehensively evaluating the feature similarity of multiple abnormal channels, it is possible to effectively distinguish whether batch anomalies are caused by sample program mismatch or systemic equipment failure, and thus take targeted handling strategies: the former provides early warning and leaves the decision to the experimenters, preserving the flexibility of the experiment; the latter immediately stops the experiment to protect the samples and equipment, enhancing the security of the system.

[0015] A second aspect of this invention provides an adaptive control system for a multi-channel electrostatic converter, the system comprising: The benchmark determination module is used to acquire descriptive data of the cell suspension to be processed and determine the corresponding diagnostic model based on the descriptive data. The monitoring module is used to acquire the electrical data of each working channel of the electro-particle converter in real time. The preprocessing module is used to extract features from the electrical data of the target channel to obtain a feature vector; the target channel is any working channel of the electro-particle converter. The identification module is used to input the feature vector into the diagnostic model to obtain the detection result of the target channel; The judgment module is used to make a comprehensive judgment based on the detection results of multiple channels to determine the working status; the working status includes normal, single point abnormality, and batch abnormality. The control module is used to execute corresponding processing strategies based on the working status; the processing strategies include correcting abnormal channels and closing abnormal channels.

[0016] Optionally, one diagnostic model corresponds to one type of cell suspension; the construction process of each diagnostic model includes: Acquire electroporation feedback data to identify normal clusters in the target cell type suspension; the electroporation feedback data is input by the user after the experiment; the normal clusters include data records with multiple electroporation results showing normality; For each normal data record, feature extraction is performed to obtain a sample feature vector; Based on the sample feature vectors of multiple data records, a single-class support vector machine is trained to obtain the diagnostic model corresponding to the target type of cell suspension.

[0017] Optionally, the control module includes: The single-point control module is used to determine the compensation of each abnormal channel if the working state is a single-point abnormality; if it can be compensated, then compensation is performed; otherwise, the abnormal channel that cannot be compensated is closed. The batch control module is used to determine the consistency of all abnormal channels if the working status is batch abnormal; if they are consistent, a warning message will pop up and the system will determine whether to continue the electronic transfer operation based on the feedback; otherwise, all channels will be shut down.

[0018] Optionally, the single-point control module includes: The reference determination module is used to calculate the average value of the current waveform of the normal channel in the current batch to obtain the reference waveform; The waveform similarity calculation module is used to calculate the similarity between the current waveform of the target abnormal channel and the reference waveform to obtain the waveform similarity; the target abnormal channel is any one of the channels identified as abnormal. The compensability determination module is used to determine that the target abnormal channel is compensable if the waveform similarity is greater than a preset stability threshold. The compensation calculation module is used to calculate the compensation coefficient based on the average steady-state current of the normal channel and the steady-state current value of the target abnormal channel, and to use the compensation coefficient to compensate the pulse voltage. The compensation execution module is used to apply a compensated pulse to the target abnormal channel starting from the next pulse.

[0019] Optionally, the batch control module includes: The feature similarity calculation module is used to calculate the similarity of feature vectors of any two abnormal channels, obtain multiple sets of similarity data, and calculate the average value of the multiple sets of similarity data to obtain the overall similarity. The consistency determination module is used to determine that all abnormal channels have consistent features if the overall similarity is greater than a preset consistency threshold. Attached Figure Description

[0020] Figure 1 A flowchart illustrating an adaptive control method for a multi-channel electro-particle transducer provided in an embodiment of the present invention; Figure 2 This is an architecture diagram of an adaptive control system for a multi-channel electro-particle transducer provided in an embodiment of the present invention. Detailed Implementation

[0021] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features and effects of the present invention is provided in conjunction with the accompanying drawings and preferred embodiments.

[0022] This invention provides an adaptive control method for a multi-channel electro-particle transducer. See also... Figure 1 , Figure 1 This is a flowchart illustrating an adaptive control method for a multi-channel electro-particle transducer, provided as an embodiment of the present invention. The method includes the following steps.

[0023] S1. Obtain descriptive data of the cell suspension to be processed, and determine the corresponding diagnostic model based on the descriptive data.

[0024] S2 acquires electrical data for each working channel of the electro-particle converter in real time.

[0025] S3. Feature extraction is performed on the electrical data of the target channel to obtain the feature vector. The target channel is any working channel of the electro-particle converter.

[0026] S4. Input the feature vector into the diagnostic model to obtain the detection result of the target channel.

[0027] S5 makes a comprehensive judgment based on the detection results of multiple channels to determine the working status.

[0028] S6 executes the corresponding processing strategy based on the working status.

[0029] The descriptive data includes cell type identifiers and invoked preset program identifiers; electrical data includes voltage and current waveform data during the application of the electrical pulse. The feature vector includes multiple features extracted from the current waveform: peak current, rise time, fall time, steady-state current value, waveform area, etc. Detection results include normal and abnormal. Operating status includes normal, single-point abnormality, and batch abnormality. Handling strategies include correcting abnormal channels and shutting down abnormal channels.

[0030] An adaptive control method for a multi-channel electroporator provided by this invention collects electrical data from each channel in real time during electroporation and performs online analysis of the channel status using feature extraction and diagnostic models. This enables real-time diagnosis of the status of individual channels and the entire batch, allowing for timely identification and differentiation between normal and abnormal channels during the experiment. Compared to existing technologies that rely on endpoint detection results, this invention can correct or shut down abnormal channels based on diagnostic results, effectively preventing the continuation of invalid electroporation, reducing waste of cell samples and reagents, and improving the intelligence level of multi-channel electroporation experiments.

[0031] In one implementation, pulses are applied sequentially to cell suspensions of the same batch on a multichannel electroporator, real-time anomaly detection is performed during the electroporation of the first group of cell suspensions, and the processing results of the channels are applied to the electroporation process of subsequent groups.

[0032] In one embodiment, the user can conduct experiments by controlling variables, in which case the electroporation system does not perform sample feedback anomaly checks. After determining the optimal control parameters for a certain type of cell sample, the user can fix them as the system's preset program, which can then be directly called in subsequent transfections for batch transfections. When calling the preset program for batch transfections, anomaly checks are performed to prevent sample waste.

[0033] In one embodiment, a type of cell suspension corresponds to an electroporation procedure and a diagnostic model. The diagnostic model is a classification support vector machine. The process of constructing the diagnostic model includes: Data Acquisition: After each experiment, users upload feedback data via the software client, including the conversion rate and viability of various types of cell suspensions after electroporation, and mark qualified data records as normal data records. For any type of cell suspension, designated as the target type cell suspension, multiple normal data records are collected to obtain its normal cluster.

[0034] Sample feature extraction: Extract features from each normal data record to obtain a sample feature vector.

[0035] Training: Based on the sample feature vectors of multiple data records, a single-class support vector machine is trained to obtain the diagnostic model corresponding to the target type of cell suspension.

[0036] This embodiment constructs a corresponding one-class support vector machine diagnostic model for each type of cell suspension, and trains it using only normal data with qualified conversion and survival rates after the experiment. This enables the model to accurately characterize the electrical characteristic distribution of this type of cell under normal electroporation conditions without relying on a large number of abnormal samples, thereby improving the accuracy and feasibility of anomaly detection. At the same time, the normal clusters formed by multiple normal data enhance the robustness of the model to fluctuations in operating conditions, providing a stable and reliable basis for real-time diagnosis and adaptive control of multi-channel electroporators.

[0037] In one embodiment, step S5, determining the working status by comprehensively judging the detection results of multiple channels, includes: If all channels are not functioning properly, the electro-particle converter is considered to be in normal working condition.

[0038] The percentage of abnormal channels to the total number of channels is denoted as the abnormality ratio.

[0039] If the anomaly rate is less than the first preset threshold (e.g., 30%), the working status is determined to be a single point of failure.

[0040] If the anomaly rate is greater than the second preset threshold (e.g., 80%), the working status is determined to be batch anomaly.

[0041] If the abnormality rate is between the first preset threshold and the second preset threshold, it is determined to be an unknown abnormality, and a warning message will pop up, allowing the user to decide whether to continue.

[0042] In one embodiment, step S6, executing the corresponding processing strategy based on the working state, includes: Scenario 1: If the working status is a single point of failure, then each abnormal channel is assessed for compensation; if it can be compensated, then compensation is performed; otherwise, the abnormal channel that cannot be compensated is closed.

[0043] Scenario 2: If the working status is batch abnormal, then perform a consistency check on all abnormal channels; if they are consistent, then pop up a warning message and determine whether to continue the electronic transfer operation based on the feedback; otherwise, close all channels.

[0044] In one implementation, the specific handling strategy for scenario 1 includes: Step 1: Calculate the average value of the current waveform of the normal channel in the current batch to obtain the reference waveform.

[0045] Step two: For any abnormal channel, denoted as the target abnormal channel, calculate the similarity between the current waveform of the target abnormal channel and the reference waveform to obtain the waveform similarity. Specifically, the Pearson correlation coefficient can be used.

[0046] Step 3: If the waveform similarity is greater than the preset stability threshold (e.g., 90%), the target abnormal channel is determined to be compensable, and proceed to step 4; otherwise, the target abnormal channel is determined to be uncompensable, and the target abnormal channel is closed.

[0047] Step 4: Based on the average steady-state current I of the normal channel avg and the steady-state current value I of the target abnormal channel abnormal Calculate the compensation coefficient k comp The pulse voltage is compensated using a compensation coefficient; the formula is: Among them, V set It is the voltage preset in the electric rotation program; V max It is the maximum allowable voltage of the electric transfer system; the min function represents taking the smaller of the two; V comp This is the compensated voltage.

[0048] Step 5: Starting from the next pulse, apply the compensated pulse to the target abnormal channel.

[0049] Step 6: If the channel is still abnormal after compensation, then close the channel.

[0050] This implementation method uses the average current waveform of the normal channels in the current batch as a reference, which can dynamically reflect the real-time operating conditions and avoid the deviation caused by using fixed parameters. By comparing waveform similarity with the stability threshold, it can accurately distinguish between compensable and uncompensable anomalies, preventing incorrect correction of severely abnormal channels. On this basis, the compensation coefficient is calculated based on the steady-state current and the pulse voltage is adaptively adjusted, so that the electrical response of the abnormal channel can quickly return to the normal level, thereby improving the stability of the overall electrical conversion process and the channel utilization rate without affecting the operation of other channels.

[0051] In one implementation, the specific processing strategy for scenario 2 includes: Step one: Calculate the similarity between the feature vectors of any two abnormal channels to obtain multiple sets of similarity data. Specifically, cosine similarity can be used.

[0052] Step 2: Calculate the average of multiple sets of similarity data to obtain the overall similarity.

[0053] Step 3: If the overall similarity is greater than the preset consistency threshold (e.g., 85%), all abnormal channels are determined to have consistent features.

[0054] Step four: If they match, a warning message will pop up on the software client to inform the user that the current batch of samples is incompatible with the called program. The software will respond to the user's instructions to continue, interrupt, or correct the electrotransfer operation of subsequent samples.

[0055] Step 5: If there is no consistency, close all channels, stop the electroporation operation of subsequent group samples, and pop up an error message, which may indicate a systemic failure of the electroporator.

[0056] This implementation method achieves effective differentiation of the causes of anomalies by quantitatively analyzing the characteristic consistency of the abnormal channels in a batch: when the overall similarity is higher than the consistency threshold, it indicates that the anomaly is consistent and is more likely caused by a mismatch between the sample type and the calling program. By popping up an alert on the client and introducing manual decision-making, misjudgment as equipment failure is avoided, and the continuity and flexibility of experiments are improved. When the anomaly is inconsistent, it is judged as a potential systemic failure and all channels are shut down to prevent sample waste or equipment damage, thereby improving the overall safety and reliability of the multi-channel electroporator.

[0057] This invention provides an adaptive control system for a multi-channel electrostatic precipitator. See also... Figure 2 , Figure 2 A flowchart illustrating an adaptive control system for a multi-channel electrostatic precipitator, provided as an embodiment of the present invention. The system includes: The baseline determination module is used to acquire descriptive data of the cell suspension to be processed and determine the corresponding diagnostic model based on the descriptive data. The monitoring module is used to acquire the electrical data of each working channel of the electro-particle converter in real time. The preprocessing module is used to extract features from the electrical data of the target channel to obtain feature vectors; The recognition module is used to input feature vectors into the diagnostic model to obtain the detection results of the target channel; The judgment module is used to make a comprehensive judgment based on the detection results of multiple channels to determine the working status; The control module is used to execute corresponding processing strategies based on the working status.

[0058] An adaptive control system for a multi-channel electroporator provided by this invention collects electrical data from each channel in real time during electroporation and performs online analysis of the channel status using feature extraction and diagnostic models. This enables real-time diagnosis of the status of individual channels and the entire batch, allowing for timely identification and differentiation between normal and abnormal channels during the experiment. Compared to existing technologies that rely on endpoint detection results, this invention can correct or shut down abnormal channels based on diagnostic results, effectively preventing the continuation of invalid electroporation, reducing waste of cell samples and reagents, and improving the intelligence level of multi-channel electroporation experiments.

[0059] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention are within the scope of the claims of the present invention.

Claims

1. An adaptive control method for a multi-channel electro-particle transducer, characterized in that, The method includes: Obtain descriptive data of the cell suspension to be processed, and determine the corresponding diagnostic model based on the descriptive data; Real-time acquisition of electrical data for each working channel of the electro-particle converter; Feature vectors are obtained by extracting features from the electrical data of the target channel; the target channel is any working channel of the electro-particle converter. The feature vector is input into the diagnostic model to obtain the detection result of the target channel; The working status is determined by comprehensively judging the test results from multiple channels; the working status includes normal, single point abnormality, and batch abnormality. Based on the operating status, execute the corresponding processing strategy; the processing strategy includes correcting abnormal channels and closing abnormal channels.

2. The adaptive control method for a multi-channel electro-particle transducer according to claim 1, characterized in that, One type of cell suspension corresponds to one diagnostic model; the construction process of each diagnostic model includes: Acquire electroporation feedback data to identify normal clusters in the target cell type suspension; the electroporation feedback data is input by the user after the experiment; the normal clusters include data records with multiple electroporation results showing normality; For each normal data record, feature extraction is performed to obtain a sample feature vector; Based on the sample feature vectors of multiple data records, a single-class support vector machine is trained to obtain the diagnostic model corresponding to the target type of cell suspension.

3. The adaptive control method for a multi-channel electro-particle transducer according to claim 1, characterized in that, The step of executing the corresponding processing strategy based on the working state includes: If the working status is a single point of failure, then compensation is determined for each abnormal channel; if it can be compensated, then compensation is performed; otherwise, the abnormal channel that cannot be compensated is closed. If the working status is batch abnormal, then a consistency check is performed on all abnormal channels; if they are consistent, a warning message is displayed and a decision is made on whether to continue the electronic transfer operation based on the feedback; otherwise, all channels are shut down.

4. The adaptive control method for a multi-channel electro-particle transducer according to claim 3, characterized in that, The compensation process for each abnormal channel includes: Based on the normal channels of the current batch, calculate the average value of its current waveform to obtain the reference waveform; A similarity calculation is performed between the current waveform of the target abnormal channel and the reference waveform to obtain the waveform similarity; the target abnormal channel is any one of the channels identified as abnormal. If the waveform similarity is greater than a preset stability threshold, then the target abnormal channel is determined to be compensable. The compensation coefficient is calculated based on the average steady-state current of the normal channel and the steady-state current of the target abnormal channel, and the compensation coefficient is used to compensate the pulse voltage. Starting from the next pulse, a compensated pulse is applied to the target abnormal channel.

5. The adaptive control method for a multi-channel electro-particle transducer according to claim 3, characterized in that, Consistency checks on all abnormal channels include: Calculate the similarity of feature vectors between any two abnormal channels to obtain multiple sets of similarity data; Calculate the average value of the multiple sets of similarity data to obtain the overall similarity. If the overall similarity is greater than the preset consistency threshold, all abnormal channels are determined to have consistent features.

6. An adaptive control system for a multi-channel electro-particle transducer, characterized in that, The system includes: The benchmark determination module is used to acquire descriptive data of the cell suspension to be processed and determine the corresponding diagnostic model based on the descriptive data. The monitoring module is used to acquire the electrical data of each working channel of the electro-particle converter in real time. The preprocessing module is used to extract features from the electrical data of the target channel to obtain a feature vector; the target channel is any working channel of the electro-particle converter. The identification module is used to input the feature vector into the diagnostic model to obtain the detection result of the target channel; The judgment module is used to make a comprehensive judgment based on the detection results of multiple channels to determine the working status; the working status includes normal, single point abnormality, and batch abnormality. The control module is used to execute corresponding processing strategies based on the working status; the processing strategies include correcting abnormal channels and closing abnormal channels.

7. An adaptive control system for a multi-channel electro-particle transducer according to claim 6, characterized in that, One type of cell suspension corresponds to one diagnostic model; the construction process of each diagnostic model includes: Acquire electroporation feedback data to identify normal clusters in the target cell type suspension; the electroporation feedback data is input by the user after the experiment; the normal clusters include data records with multiple electroporation results showing normality; For each normal data record, feature extraction is performed to obtain a sample feature vector; Based on the sample feature vectors of multiple data records, a single-class support vector machine is trained to obtain the diagnostic model corresponding to the target type of cell suspension.

8. An adaptive control system for a multi-channel electro-particle transducer according to claim 6, characterized in that, The control module includes: The single-point control module is used to determine the compensation of each abnormal channel if the working state is a single-point abnormality; if it can be compensated, then compensation is performed; otherwise, the abnormal channel that cannot be compensated is closed. The batch control module is used to determine the consistency of all abnormal channels if the working status is batch abnormal; if they are consistent, a warning message will pop up and the system will determine whether to continue the electronic transfer operation based on the feedback; otherwise, all channels will be shut down.

9. An adaptive control system for a multi-channel electro-particle transducer according to claim 8, characterized in that, The single-point control module includes: The reference determination module is used to calculate the average value of the current waveform of the normal channel in the current batch to obtain the reference waveform; The waveform similarity calculation module is used to calculate the similarity between the current waveform of the target abnormal channel and the reference waveform to obtain the waveform similarity; the target abnormal channel is any one of the channels identified as abnormal. The compensability determination module is used to determine that the target abnormal channel is compensable if the waveform similarity is greater than a preset stability threshold. The compensation calculation module is used to calculate the compensation coefficient based on the average steady-state current of the normal channel and the steady-state current value of the target abnormal channel, and to use the compensation coefficient to compensate the pulse voltage. The compensation execution module is used to apply a compensated pulse to the target abnormal channel starting from the next pulse.

10. An adaptive control system for a multi-channel electro-particle transducer according to claim 8, characterized in that, The batch control module includes: The feature similarity calculation module is used to calculate the similarity of feature vectors of any two abnormal channels, obtain multiple sets of similarity data, and calculate the average value of the multiple sets of similarity data to obtain the overall similarity. The consistency determination module is used to determine that all abnormal channels have consistent features if the overall similarity is greater than a preset consistency threshold.

Images