Air conditioner control method and system based on delay angle control and ai harmonic fingerprint identification

Abstract

The application relates to the field of intelligent air conditioner control, and provides an air conditioner control method and system based on delay angle control and AI harmonic fingerprint identification, which comprises the following steps: determining a voltage zero-crossing time according to a power grid frequency, combining an initial setting instruction and a remedial control instruction to generate a theoretical delay angle, obtaining a trigger offset amount to drive a switching device, and forming a carrier voltage waveform; determining a local zero-crossing time and a waveform distortion point from the carrier voltage waveform, solving an actual analysis angle, determining a restored control semantic, and generating an air conditioner control interface signal; collecting a controlled current and a voltage sampling sequence, extracting a harmonic feature set to construct a harmonic fingerprint vector, and combining an angle constraint interval to generate an actual operation matrix; calculating a consistency confidence degree from the initial setting instruction and the actual operation matrix, combining the harmonic feature set to obtain an abnormal type mark, and generating a remedial control flow. The application combines delay angle waveform modulation and AI harmonic fingerprint working condition mapping, and constructs an intelligent analysis and abnormal self-recovery closed loop of the air conditioner control system.

Description

Technical Field

[0001] This invention relates to the field of intelligent air conditioning control, and in particular to an air conditioning control method and system based on delay angle control and AI harmonic fingerprint recognition. Background Technology

[0002] With the increasing demand for energy-saving renovations and centralized management of large public buildings, existing air conditioning systems generally face problems such as limited physical wiring, unstable signal transmission, and insufficient feedback capabilities in remote control and status monitoring. Existing wired or wireless control methods struggle to balance renovation costs with transmission reliability. Therefore, how to combine delay angle control and harmonic characteristic recognition to achieve air conditioning control, status identification, and anomaly closed-loop compensation has become a pressing technical problem to be solved in this field.

[0003] However, current technologies still face numerous challenges. In energy-saving retrofit scenarios for existing air conditioning systems in large public buildings, existing control links often struggle to reliably transmit control commands due to limited cabling conditions in the equipment room, complex shielding environments, and strong interference from variable frequency loads. In such cases, relying solely on traditional wired or ordinary wireless methods for air conditioning control may result in devices failing to execute commands as expected, while the management end cannot obtain timely and effective feedback. Furthermore, existing status monitoring solutions typically rely solely on the presence or absence of current or changes in switching quantities to determine whether the air conditioner is on, making it difficult to identify false operating or semi-fault states such as compressor start-stop failures or abnormal four-way valve reversals—conditions where the system is powered on but not effectively cooling or heating. If the system misjudges such anomalies as normal operation, timely remedial control or fault intervention is impossible, leading not only to continuous inefficient operation of the air conditioner and reduced energy-saving retrofit effects, but also to potential failures in regional temperature control, increased abnormal equipment wear, and ultimately affecting the stable operation and centralized monitoring effectiveness of the building's air conditioning system. Summary of the Invention

[0004] To achieve the above objectives, this invention provides an air conditioning control method based on delay angle control and AI harmonic fingerprint recognition, the specific technical solution of which is as follows:

[0005] The zero-crossing reference time of the voltage is determined based on the real-time frequency of the power grid. The theoretical delay angle is generated by combining the initial setting command and the remedial control command. The trigger time offset is obtained based on the theoretical delay angle, and the power switching device is driven to conduct to form a carrier voltage waveform.

[0006] The local zero-crossing reference time and waveform distortion point are determined based on the carrier voltage waveform to calculate the actual resolution angle. The restored control semantics are determined based on the actual resolution angle, and the air conditioning control interface signal is generated based on the restored control semantics.

[0007] Collect the controlled current sampling sequence and controlled voltage sampling sequence after the air conditioner control interface signal is triggered, extract the harmonic feature set and construct the harmonic fingerprint feature vector, and generate the actual operating state matrix by combining the angle constraint interval corresponding to the control action;

[0008] The consistency confidence level is calculated based on the initial setting instructions and the actual operating state matrix. An anomaly type flag is obtained based on the consistency confidence level and harmonic feature set. A remedial control flow is then generated based on the anomaly type flag for feedback.

[0009] Furthermore, the method for forming the carrier voltage waveform includes:

[0010] The half-cycle search window is determined based on the real-time frequency of the power grid. Curvature discrimination processing is performed on the collected voltage sampling sequence to determine the zero-crossing candidate time. The voltage zero-crossing reference time is generated by combining the time delay compensation of the sampling link.

[0011] A composite control quantity is generated based on the initial setting command and the remedial control command, and the composite control quantity is mapped to the theoretical delay angle through a preset mapping function;

[0012] The power grid cycle duration is determined based on the real-time frequency of the power grid, and the trigger time offset is obtained by combining the theoretical delay angle and the link compensation time.

[0013] Timing begins based on the voltage zero-crossing reference moment. When the timing reaches the trigger time offset, the power switching device is driven to conduct, forming a carrier voltage waveform.

[0014] Furthermore, the method for performing curvature discrimination processing includes:

[0015] The half-cycle reference duration is determined based on the real-time frequency of the power grid, and a half-cycle search window is constructed based on the half-cycle reference duration.

[0016] Within the half-cycle search window, sampling points with voltage values ​​close to zero are selected from the voltage sampling sequence as candidate points; the second-order difference is calculated based on the relationship between the changes of adjacent sampling values ​​before and after each candidate point to determine the zero-crossing candidate time.

[0017] Furthermore, the method for generating the air conditioning control interface signal includes:

[0018] A voltage sampling sequence is formed based on the carrier voltage waveform. The local zero-crossing reference time is determined by polarity determination, and the waveform distortion points are extracted based on the second-order difference of the voltage sampling sequence.

[0019] Calculate the time difference between the waveform distortion point and the local zero-crossing reference time, and convert the time difference and half-cycle reference duration to obtain the actual resolution angle;

[0020] The actual resolution angle is superimposed with a preset angle offset correction amount to obtain the corrected resolution angle, and the restoration control semantics are determined according to the target angle interval to which the corrected resolution angle belongs in the preset angle interval table.

[0021] Based on the restored control semantics, the corresponding control protocol item is retrieved from the local protocol library and combined with the local zero-crossing reference time to generate the air conditioning control interface signal, so that the controlled air conditioning terminal can perform the corresponding control action.

[0022] Furthermore, the method for determining polarity includes:

[0023] Based on the carrier voltage waveform, a voltage sampling sequence arranged in chronological order is formed; for two adjacent sample values ​​in the voltage sampling sequence, it is determined whether their signs have changed;

[0024] When two adjacent sampled values ​​are one positive and one negative, or the previous sampled value is zero and the next sampled value is non-zero, or the previous sampled value is non-zero and the next sampled value is zero, it is determined that there is a zero-crossing position in the adjacent sampling interval.

[0025] The zero-crossing position is determined as the local zero-crossing reference time.

[0026] Furthermore, the method for generating the actual operating state matrix includes:

[0027] The controlled current sampling sequence and controlled voltage sampling sequence are collected after the air conditioner control interface signal is triggered. A fast Fourier transform is performed on the controlled current sampling sequence to obtain the effective value of each harmonic current. The phase offset of each harmonic is extracted by combining the controlled voltage sampling sequence, and the total harmonic distortion rate is calculated to form a harmonic feature set.

[0028] The target harmonic order set is selected based on the harmonic feature set and the total harmonic distortion rate. The effective values ​​of each harmonic current in the target harmonic order set are converted by ratio to obtain the normalized harmonic amplitude. The normalized harmonic amplitude, phase offset and total harmonic distortion rate are combined in a preset order to construct the harmonic fingerprint feature vector.

[0029] The angle constraint interval corresponding to the control action is determined based on the harmonic fingerprint feature vector and the controlled voltage sampling sequence. The harmonic feature components located within the angle constraint interval in the harmonic fingerprint feature vector are combined into a joint feature, and the actual operating state matrix is ​​generated according to the state mapping rule.

[0030] Furthermore, the method for calculating the total harmonic distortion rate includes:

[0031] Collect the controlled current sampling sequence after the air conditioner control interface signal is triggered; perform a fast Fourier transform on the controlled current sampling sequence to obtain the effective value of the fundamental current and the effective values ​​of the second to Nth harmonic currents;

[0032] The fundamental current RMS value is the RMS current value extracted at the frequency point corresponding to the power supply reference frequency; the power supply reference frequency is the reference frequency of the AC power supply connected to the current controlled air conditioning circuit.

[0033] The effective values ​​of the 2nd to Nth harmonic currents are squared and summed, and then the sum is squared to obtain the composite effective value.

[0034] Divide the synthesized effective value by the effective value of the fundamental current and multiply by 100% to obtain the total harmonic distortion rate.

[0035] Furthermore, the method for generating the remedial control flow includes:

[0036] The initial setting command and the actual running state matrix are obtained. The initial setting command is encoded into a target state vector. The actual state vector is extracted from the actual running state matrix. The state harmonic potential energy is calculated based on the target state vector and the actual state vector to obtain the consistency confidence.

[0037] The evolutionary gradient vector is calculated based on the consistency confidence level. The evolutionary gradient vector and harmonic feature set are jointly discriminated to obtain the anomaly type indicator.

[0038] Based on the anomaly type flag, the delay angle control parameters are reorganized to generate a remedial control command, which is then encapsulated into a remedial control flow for feedback.

[0039] Furthermore, the joint discrimination method includes:

[0040] When the consistency confidence is less than the preset consistency threshold, the evolution gradient vector is obtained based on the change in consistency confidence during the current observation period, the corresponding time interval, and the evolution correction coefficient.

[0041] The observation time period corresponding to the consistency confidence is matched with the sampling time period corresponding to the harmonic feature set. When the magnitude of the evolution gradient vector is greater than the preset mutation threshold and the harmonic components of each order in the harmonic feature set do not meet the effective identification conditions, the abnormality type flag is determined as a control signal transmission abnormality.

[0042] If the magnitude of the evolution gradient vector is not greater than a preset mutation threshold, and there are harmonic components in the harmonic feature set that correspond to the target control action, but the amplitude deviation, phase shift, or inter-order amplitude ratio of the harmonic components relative to the preset reference harmonic features exceeds a preset deviation range, the abnormality type flag will be determined as an abnormal action of the execution component.

[0043] Furthermore, the method for parameter recombination includes:

[0044] When the anomaly type is marked as a control signal transmission anomaly, the actual resolution angle is retrieved, and the difference between the actual resolution angle and the target delay angle corresponding to the current control task is calculated to obtain the residual phase deviation.

[0045] Based on the residual phase deviation, the target delay angle of the next control cycle is compensated to obtain the optimized delay angle, and the optimized delay angle is written into the angle parameter field of the delay angle control command to generate a remedial control command.

[0046] When the anomaly type flag indicates an abnormal action of the executing component, additional corrections are performed on the delay angle conduction interval in the remedial control command to adjust the start or end point of the delay angle and generate a remedial control command.

[0047] The remedial control command is encapsulated into a remedial control flow according to a preset data format and fed back to the sending entry.

[0048] An air conditioning control system based on delay angle control and AI harmonic fingerprint recognition is used to implement the above-mentioned air conditioning control method based on delay angle control and AI harmonic fingerprint recognition. The system includes a carrier voltage modulation module, an instruction parsing module, a status monitoring module, and an anomaly diagnosis and compensation module.

[0049] The carrier voltage modulation module is used to determine the voltage zero-crossing reference time based on the real-time frequency of the power grid, generate a theoretical delay angle by combining the initial setting command and the remedial control command, obtain the trigger time offset based on the theoretical delay angle, and drive the power switching device to conduct to form a carrier voltage waveform.

[0050] The instruction parsing module is used to determine the local zero-crossing reference time and waveform distortion point based on the carrier voltage waveform, to calculate the actual resolution angle, to determine the restored control semantics based on the actual resolution angle, and to generate the air conditioning control interface signal based on the restored control semantics.

[0051] The status monitoring module is used to collect the controlled current sampling sequence and controlled voltage sampling sequence after the air conditioning control interface signal is triggered, extract the harmonic feature set and construct the harmonic fingerprint feature vector, and generate the actual operating status matrix by combining the angle constraint interval corresponding to the control action.

[0052] The anomaly diagnosis and compensation module is used to calculate the consistency confidence level based on the initial setting command and the actual operating state matrix, obtain the anomaly type flag based on the consistency confidence level and harmonic feature set, and generate a remedial control flow for feedback based on the anomaly type flag.

[0053] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0054] This invention achieves temporal alignment of images and audio through multimodal synchronous acquisition and standardized processing, supports joint discrimination of visual and acoustic features, reduces misjudgments caused by modal asynchrony, and enables rapid localization and response to multimodal abnormal behaviors. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 This is a flowchart illustrating the principle of the air conditioning control method based on delay angle control and AI harmonic fingerprint recognition of the present invention.

[0058] Figure 2 This is a functional block diagram of the air conditioning control system based on delay angle control and AI harmonic fingerprint recognition of the present invention. Detailed Implementation

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

[0060] Example 1:

[0061] Please see Figure 1 As shown, this embodiment provides an air conditioning control method based on delay angle control and AI harmonic fingerprint recognition, including:

[0062] Step S1000, based on the real-time frequency of the power grid Determine the reference time for voltage zero crossing Combined with initial setting instructions and remedial control instructions Generate theoretical delay angle From the theoretical delay angle Calculate the trigger time offset This drives the power switching devices to conduct, forming a carrier voltage waveform. .

[0063] Specifically, this step aims to utilize the real-time frequency of the power grid. As a benchmark for time-series evolution, the initial setting instructions for discrete distributions will be used. and remedial control commands that characterize feedback deviations This is integrated into a theoretical delay angle with physical meaning. Abstract composite control quantities Converted to carrier voltage waveform The delay angle in the power transmission layer establishes a one-way information carrier with the air conditioning terminal.

[0064] Further, step S1000 includes:

[0065] Step S1100, based on the real-time frequency of the power grid A half-cycle search window is determined, and curvature discrimination processing is performed on the acquired voltage sampling sequence to determine the candidate zero-crossing time, which is then combined with the time delay compensation of the sampling link. Generating voltage zero-crossing reference time .

[0066] Specifically, this step aims to utilize the real-time frequency of the power grid acquired in real time. A time reference is established within the air conditioner power supply circuit, and instantaneous voltage signals are captured. At the physical level, the candidate zero-crossing moments are mapped to the voltage zero-crossing reference moments. This provides a time reference for the subsequent injection of delay angles under complex variable frequency load environments.

[0067] In practical implementation, in energy-saving renovation scenarios of large public buildings, air conditioning loads commonly include variable frequency compressors, fan drives, and switching power supply components. When these loads are connected, they generate instantaneous jump signals and reciprocating signals near the zero point in the power supply circuit voltage signal. If only the voltage comparator is used to detect positive and negative voltage jumps, abnormal change points near the zero point caused by load switching can easily be misjudged as zero-crossing points, leading to phase shifts in subsequent delay angle encoding. Therefore, this step does not directly use the sign change of a single sampling point as the judgment criterion, but instead first uses the actual monitored real-time frequency of the power grid. The half-cycle reference duration of the current power grid is dynamically calculated. The half-cycle reference duration This represents the theoretical time length between two consecutive zero crossings with opposite signs, which is numerically equal to the real-time frequency of the power grid. The reciprocal of twice is used to limit the half-cycle search window, determining the moment when the voltage crosses from a positive value to a negative value or from a negative value to a positive value only within the time interval that conforms to the current power grid cycle, thereby excluding anomalies that do not conform to the half-cycle evolution law from the judgment range.

[0068] During the zero-crossing identification process, this step involves the instantaneous voltage signal. Continuous sampling is performed, and the instantaneous voltage signal is... The formed voltage sampling sequence undergoes curvature discrimination processing. The specific execution logic of the curvature discrimination processing is as follows: Based on a half-cycle reference duration... Within a half-cycle search window of length , sampling points with voltage values ​​close to zero are first selected from the voltage sampling sequence as candidate points. Then, a second-order difference is calculated based on the relationship between adjacent sampling values ​​before and after each candidate point to identify candidate zero-crossing moments when the voltage transitions from the positive half-cycle to the negative half-cycle, or from the negative half-cycle to the positive half-cycle. In this way, sampling points with voltage values ​​close to zero but whose direction of change has not continuously reversed are eliminated, and only candidate zero-crossing moments that meet the condition of continuous polarity reversal are retained.

[0069] After obtaining the candidate zero-crossing moments, the corresponding instantaneous voltage signal is selected from the candidate zero-crossing moments. The moment whose absolute value is closest to zero is selected as the minimum candidate moment. Building upon this, this step further introduces the time delay compensation amount for the sampling link. To generate the voltage zero-crossing reference time The voltage zero-crossing reference time. Numerically equal to the minimum candidate time Delay compensation amount of sampling link sum.

[0070] Among them, the delay compensation amount of the sampling link This is a fixed-time compensation value obtained through reference voltage measurement during the installation and commissioning phase. It is used to compensate for the overall delay caused by voltage sampling, circuit filtering, analog-to-digital conversion, and program processing. The reference voltage measurement refers to simultaneously acquiring two zero-crossing moments during the measurement period after the control device is connected to the air conditioner power supply circuit, when the air conditioner actuators are not performing start / stop or switching actions. One is the sampled zero-crossing moment output by the voltage sampling branch, and the other is the reference zero-crossing moment obtained by the reference measurement branch detecting the voltage of the same power supply circuit. The sampled zero-crossing moment and the reference zero-crossing moment are compared to obtain the time difference between them. This time difference is written into the control device as a fixed compensation parameter, i.e., the delay compensation amount. The measurement period during which the air conditioning actuators did not perform start-stop or switching actions refers to the period during which the compressor, four-way valve, and indoor and outdoor fans did not receive switching commands.

[0071] For example, in a shopping mall's central air conditioning circuit, when multiple inverter devices start simultaneously, local fluctuations often occur near the zero point of the voltage waveform. If the zero-crossing point is determined directly based on the positive and negative jumps of the voltage comparator output, the abnormal sampling point corresponding to the local fluctuation may be identified as a zero-crossing point, causing the subsequent 5° start-up command to be interpreted as a 10° mode switching command. Using this method, this step first relies on the real-time monitored power grid frequency. Provide the current power grid half-cycle reference duration. Then, within this half-cycle search window, search only for the smallest candidate moment that satisfies the continuous change in polarity. Finally, the delay compensation amount of the sampling link is added. Obtain the voltage zero-crossing reference time The zero-crossing reference time of the voltage obtained in this way It is not determined by a single abnormal sampling point, but by a combination of period constraints, waveform constraints, and delay compensation. Therefore, the subsequent theoretical delay angle... The injection position will not be misaligned due to local burrs, and the air conditioning control command can stably fall on the predetermined phase.

[0072] Step S1200, according to the initial setting command and remedial control instructions Generate composite control quantity And the composite control quantity is mapped through a preset mapping function. Mapped to theoretical delay angle .

[0073] Specifically, this step aims to follow the initial setting instructions. and remedial control instructions Generate composite control quantity Then the composite control quantity Mapped to the theoretical delay angle within half a cycle of a sine wave This allows the air conditioning control command to be mapped to a delayed trigger position within half a cycle of the AC power supply voltage, enabling the control device to perform delay angle control on the existing power supply circuit and providing corresponding waveform features for subsequent AI harmonic fingerprint recognition.

[0074] In the specific implementation process, in the scenario of centralized control of air conditioning in large public buildings, this step obtains the initial setting command for the current control cycle. and the remedial control command from the previous control cycle in the feedback loop The initial setting command, in particular... The remedial control command is used to represent the target control command of the current air conditioning equipment. This is used to represent the correction amount between the actual execution result of the previous control cycle and the target control instruction. When in the first control cycle and the actual execution result of the previous control cycle has not yet been obtained, the remedial control instruction... Take the value of zero.

[0075] Obtain the initial setting instruction and the remedial control command Then, this step performs a weighted summation of the two values ​​to obtain the composite control quantity. The composite control quantity Numerically equal to the initial setting command. With remedial control instructions The sum after multiplying by the first weighting coefficient. The first weighting coefficient is a pre-set proportional coefficient used to limit the remedial control command. Initial setting instructions The correction range is adjusted to avoid remedial control commands. Directly replace the initial setting command This can cause a jump in control semantics. Preferably, the value of the first weight coefficient is greater than or equal to 0 and less than or equal to 1.

[0076] After obtaining the composite control quantity Then, this step uses a preset mapping function to convert the composite control quantity... Mapped to theoretical delay angle The preset mapping function represents a mapping function from air conditioning control commands to delay angle intervals. It is used to map distinguishable command types within the air conditioning control commands to different delay angle intervals within a half-cycle of a sine wave, allowing different control intentions to correspond to different delay angle positions. The angle interval within the half-cycle of the sine wave is based on the voltage zero-crossing reference time in step S1100. The corresponding zero-degree position is the 0° to 180° range from the starting point.

[0077] The specific mapping process of the preset mapping function includes: based on the composite control quantity Determine the corresponding instruction type; select the corresponding delay angle segment based on the instruction type; and then, based on the composite control quantity... The theoretical delay angle is determined by the control level corresponding to this instruction type. The specific angle value within the said delay angle segment.

[0078] The instruction type refers to the control instruction type corresponding to the controlled component of the air conditioner. Different instruction types correspond to different delay angle segments, and there is a one-to-one correspondence. The instruction type includes at least one of the following: compressor start / stop instruction type, four-way valve reversing instruction type, and fan speed adjustment instruction type. The control device pre-assigns non-overlapping delay angle segments to each instruction type, so that different instruction types correspond to different delay angle segments, thereby allowing each type of instruction to correspond to a different delayed trigger position within half a cycle of the power supply voltage.

[0079] The control level represents the degree of adjustment within the same instruction type. It does not change the instruction type and is only used to determine the theoretical delay angle. The specific location within the delay angle segment corresponding to this instruction type. Therefore, when the instruction type remains unchanged, the composite control quantity... Changes are reflected in the theoretical delay angle through the control level. Movement within the same delay angle segment without crossing into delay angle segments corresponding to other instruction types.

[0080] For example, the compressor start / stop command type corresponds to the first delay angle segment, the four-way valve reversing command type corresponds to the second delay angle segment, and the fan speed command type corresponds to the third delay angle segment; when the composite control quantity When determined to be a compressor start / stop command type, the theoretical delay angle It only takes a value within the first delay angle segment; when this composite control quantity The theoretical delay angle is calculated based on the different control levels corresponding to the compressor start / stop command types. Different angle values ​​are taken within the first delay angle range to represent different degrees of adjustment under the same command type. In this way, the command type and control level respectively serve the functions of category differentiation and range positioning, thus ensuring the theoretical delay angle... It also has the functions of instruction type identification and intra-segment adjustment.

[0081] Step S1300, based on the real-time frequency of the power grid Determine the power grid cycle duration and combine it with the theoretical delay angle. Link compensation time Get the trigger time offset .

[0082] Specifically, this step aims to collect the real-time frequency of the power grid in step S1100. Establish theoretical delay angle and trigger time offset The conversion relationship between them will be used to calculate the theoretical delay angle in step S1200. Converted to the trigger time offset corresponding to the execution circuit To reduce the impact of physical fluctuations in grid frequency on the delay angle control position and ensure the same theoretical delay angle. Frequency implementation in different power grids The lower half corresponds to the same half-cycle angle position, thereby keeping the control trigger signal in the existing power supply circuit consistent and providing a stable harmonic acquisition reference for subsequent AI harmonic fingerprint recognition.

[0083] In the specific implementation process, in the scenario of energy-saving renovation of large public buildings, the power supply circuit where the air conditioning equipment is located is affected by the switching of the compressor, fan and frequency converter drive unit, and the real-time frequency of the power grid is collected in real time. It will change within the control cycle. Due to the theoretical delay angle This indicates the time relative to the voltage zero-crossing reference point. The angle position is determined by the time-counting method used by the execution circuit to actually output the trigger signal, therefore it is necessary to determine the actual trigger signal based on the currently acquired real-time frequency of the power grid. This converts angular quantities into time quantities.

[0084] Therefore, this step obtains the real-time frequency of the power grid. And establish a theoretical delay angle based on the current power grid cycle duration. Offset from trigger time The conversion relationship between them. The grid cycle duration refers to the time it takes for the current alternating current to complete a full sinusoidal cycle, and its value is determined by the real-time frequency of the grid. Decision; when the real-time frequency of the power grid As the frequency increases, the power grid cycle duration decreases; when the real-time frequency of the power grid increases... When the frequency decreases, the duration of the power grid cycle increases.

[0085] The conversion relationship is as follows: first, based on the real-time frequency of the power grid... Determine the duration of the power grid cycle corresponding to the current complete AC cycle, and then based on the theoretical delay angle. The proportion of the angle within the complete 360° cycle is used to determine the zero-crossing reference time of the self-voltage. The time elapsed from the starting point to the target trigger angle; on this basis, the link compensation time is then added. To obtain the final trigger time offset. Wherein, the trigger time offset Indicates the reference time for the self-voltage zero crossing. The time offset between the start and the output trigger signal.

[0086] Furthermore, the theoretical delay angle A correspondence is established between the grid cycle length and the grid cycle length through angular proportions. Specifically, a complete cycle angle of 360° corresponds to the current grid cycle length, and the theoretical delay angle... The proportion of the full 360° angle in the cycle is equal to the proportion of the target trigger time in the current power grid cycle duration. Therefore, when the theoretical delay angle... Once determined, the corresponding time can be calculated from the current power grid cycle duration using the same proportion. For example, when the theoretical delay angle... When the angle is 90°, since 90° is one-quarter of the full cycle angle of 360°, the target triggering time corresponds to one-quarter of the current power grid cycle duration; when the theoretical delay angle is... At a 30° angle, the target trigger time corresponds to one-twelfth of the current power grid cycle length. Through this conversion, the real-time frequency of the power grid can be determined. The triggering time during the change remains at the corresponding theoretical delay angle. Location.

[0087] The link compensation time This is used to compensate for execution delays in the trigger output link, thereby correcting the actual timing of the trigger signal. The amount of time this link compensates for is... The time difference is obtained in advance during the installation and commissioning phase. Specifically, after the control device is connected to the air conditioner power supply circuit, the trigger output link outputs a trigger signal according to a preset delay time, and the time difference between the output trigger time of the control device and the actual conduction time of the actuator is detected simultaneously. The time differences measured multiple times are statistically analyzed to obtain a fixed compensation value, which is then written into the control device as the link compensation time. The actual turn-on time of the actuator can be determined by at least one of the following: the drive output detection signal, the device turn-on detection signal, or the moment of sudden change in the loop current.

[0088] Step S1400, based on the voltage zero-crossing reference time Start timing, and wait until the timer reaches the trigger time offset. At this time, the driving power switching device is turned on, forming a carrier voltage waveform. .

[0089] Specifically, this step aims to determine the voltage zero-crossing reference time based on step S1100. and the trigger time offset of step S1300 The control power switching device is turned on at the trigger moment, so that the original sinusoidal voltage waveform remains off in the initial interval after the zero crossing, and is connected to the power supply circuit at the trigger moment, thereby forming a carrier voltage waveform corresponding to the delay angle control command in the power supply circuit. The carrier voltage waveform On the one hand, it is used to transmit control commands to the controlled air conditioning terminal, and on the other hand, it is used to generate harmonic features in the power supply circuit corresponding to the control commands, so as to serve as the input basis for subsequent AI harmonic fingerprint recognition.

[0090] In the specific implementation process, during the command issuance phase of air conditioning control in large public buildings, the main station controller uses the voltage zero-crossing reference time... As the timing start point, the power switching devices in the control power supply circuit are in the off state. The off state means that the power switching devices are in the off state, and no conductive path is formed between the controlled branch circuit and the power supply circuit. During this stage, the controlled branch circuit does not conduct the corresponding first half-cycle of the original sinusoidal voltage waveform, thus keeping the output voltage within the initial interval of that half-cycle zero. The original sinusoidal voltage waveform refers to the AC voltage waveform in the power supply circuit, which is a voltage sampling sequence acquired in step S1100 in the power supply circuit without phase slicing processing.

[0091] When the timer reaches the trigger time offset At the corresponding trigger moment, the controller outputs a trigger signal, driving the power switching device to switch from the off state to the on state. The on state refers to the power switching device being in the ON state, forming a conductive path between the controlled branch circuit and the power supply circuit. After the power switching device is turned on, the controlled branch circuit receives the original sinusoidal voltage corresponding to the trigger moment, thereby causing the output voltage waveform to have a level access edge at the trigger moment. This level access edge corresponds to the theoretical delay angle. The corresponding trigger position in the time domain constitutes the waveform start characteristic of the command issuance control content in the power supply circuit. The trigger time is the conduction start time within the current half-cycle.

[0092] At the voltage zero-crossing reference time The carrier voltage waveform within the corresponding current half-cycle This can be expressed as: the reference time for the self-voltage zero crossing. From the time of voltage zero crossing reference point and trigger time offset The carrier voltage waveform before the corresponding trigger time. The value is zero; from the triggering time to the end of the current half-cycle, the carrier voltage waveform is zero. This is equal to the original sinusoidal voltage of the power supply circuit. The end time of the current half-cycle refers to the voltage zero-crossing reference time. Compared with the current half-cycle reference duration in step S1100 The sum corresponds to the time. This is achieved by adjusting the trigger time offset. It can change the carrier voltage waveform. At the start of conduction within the current half-cycle. Trigger time offset. Different, carrier voltage waveform The different positions of the access edges result in different waveform truncation methods. Different waveform truncation methods correspond to different harmonic component distributions, thus causing each control command to correspond to different harmonic characteristics in the power supply circuit.

[0093] Step S2000, based on the carrier voltage waveform Determine the local zero-crossing reference time and waveform distortion points To solve for the actual analytical angle Based on actual analytical angle Determine the reduction control semantics And based on the restoration control semantics Generate air conditioning control interface signal .

[0094] Specifically, this step aims to utilize the air conditioning terminal to analyze the carrier voltage waveform from step S1400. Real-time monitoring is performed by identifying waveform distortion points in the power supply circuit waveform. The delay angle control content in the physical carrier is restored to the control semantics. and match the local protocol library This is converted into a physical interface drive signal that directly acts on the air conditioning actuator. .

[0095] Further, step S2000 includes:

[0096] Step S2100, based on the carrier voltage waveform A voltage sampling sequence is generated, and the local zero-crossing reference time is determined by polarity determination. The waveform distortion points are extracted based on the second-order difference of the voltage sampling sequence. .

[0097] Specifically, this step aims to transform the carrier voltage waveform generated in the air conditioning power supply circuit from step S1400. As the monitoring object, the carrier voltage waveform Local sampling and waveform analysis are performed to determine the local zero-crossing reference time for local demodulation. It also identifies the position corresponding to the injection edge formed by the delay angle control, i.e., the waveform distortion point. .

[0098] In the specific implementation process, at the end-effector of the air conditioning power supply line in large public buildings, the carrier voltage waveform is... Real-time sampling is performed, and local zero-crossing reference time is executed based on the sampling results. Determine and waveform distortion points Extraction. The specific execution logic is as follows:

[0099] The end-efficiency unit acquires the carrier voltage waveform through a sampling circuit. The instantaneous voltage signal is collected and a voltage sampling sequence is formed in chronological order. Based on the voltage sampling sequence, polarity determination is first performed to identify the zero-crossing position where the voltage transitions from a positive to a negative value or from a negative to a positive value, and the identified zero-crossing position is determined as the local zero-crossing reference time. The polarity determination refers to: for two adjacent sampled values, determining whether their signs change; when two adjacent sampled values ​​are one positive and one negative, or the previous sampled value is zero and the next sampled value is non-zero, or the previous sampled value is non-zero and the next sampled value is zero, it is determined that a zero-crossing position exists within the adjacent sampling interval. The local zero-crossing reference time... Used for the voltage zero-crossing reference time in step S1400 This is to establish a half-cycle timing reference in the end-effector unit that corresponds to the delay angle control process on the master station side.

[0100] Determining the local zero-crossing reference time Then, a second-order difference operation is performed on the voltage sampling sequence to extract the carrier voltage waveform. Waveform distortion points The specific execution logic of the second-order difference operation is as follows: The adjacent differences between the sampled values ​​corresponding to the previous, current, and next time moments are subtracted again to obtain the second-order difference value corresponding to the sampling position at the current time. Here, the second-order difference value represents the degree of abrupt change in the voltage change trend at the sampling position at the current time.

[0101] The waveform distortion point The specific extraction logic is as follows: when the absolute value of the second-order difference is greater than the preset distortion judgment threshold, it is determined that there is a waveform abrupt change at the sampling position, and the corresponding position is determined as the waveform distortion point. The waveform distortion point This corresponds to the access edge position formed on the carrier voltage waveform when the power switching device switches from the off state to the on state in step S1400. Since this access edge is generated by the delay angle control action, this waveform distortion point... It can be used to characterize the trigger position of the corresponding control content.

[0102] The preset distortion judgment threshold is determined based on the baseline voltage sampling sequence collected by the end-effector when it does not receive delay angle control content. Specifically, during the time period when the power switching device remains in a constant conducting state and no phase slicing processing is performed in the power supply circuit, multiple AC half-cycle baseline voltage sampling sequences are continuously collected, and the absolute value of the second-order difference value corresponding to each sampling position is calculated. Then, based on the statistical results of the absolute values ​​of the second-order difference values, a threshold representing the upper bound of background fluctuations is determined as the preset distortion judgment threshold. The statistical results are the maximum value of the absolute value of the second-order difference value at each sampling position in the baseline voltage sampling sequences corresponding to multiple AC half-cycles.

[0103] Step S2200: Calculate waveform distortion points relative to the local zero-crossing reference time Time difference And based on the time difference and half-cycle reference duration The actual resolution angle is obtained by conversion. .

[0104] Specifically, this step is based on the local zero-crossing reference time of step S2100. and waveform distortion points and the half-cycle reference duration in step S1100 , waveform distortion points The corresponding time position is converted into the angular position within the current half-cycle of the communication to obtain the actual analytical angle. This is used to analyze the delay angle control content written by the master station into the power grid voltage waveform.

[0105] In practical implementation, in the scenario of energy-saving renovation of air conditioning in large public buildings, the end-effector receives the delay angle control content written by the master station through the power supply line. Because the power grid frequency in the power supply circuit varies, and the local sampling timing reference of the end-effector differs from the timing reference used by the master station when writing the delay angle control content, judging waveform distortion points solely based on a fixed time length becomes problematic. The timing and location can easily cause the actual analytical angle to be affected. The calculation deviation. Therefore, this step uses the half-cycle reference duration corresponding to the current AC half-cycle. As a conversion benchmark, for waveform distortion points The angle is calculated at the corresponding time position. The specific execution logic for the angle calculation is as follows:

[0106] First, obtain the local zero-crossing reference time. The local zero-crossing reference time. This is the starting moment of the current half-cycle of the exchange, used as the actual analytical angle. The zero-degree reference. Simultaneously, the half-cycle reference duration corresponding to the current half-cycle is obtained. The half-cycle reference duration This indicates the time length between two consecutive zero crossings with opposite signs, determined by the real-time frequency of the power grid. Confirmed. Then, waveform distortion points are extracted. The moment of occurrence, i.e., the moment of waveform distortion. The waveform distortion point time Indicates waveform distortion points Position within the current half-week of communication.

[0107] Obtaining the local zero-crossing reference time Waveform distortion point time and half-cycle reference duration Next, this step calculates the waveform distortion points. relative to the local zero-crossing reference time Time difference The time difference Indicates waveform distortion points The delay time relative to the current half-week starting point of the exchange.

[0108] Finally, the time difference value Based on the current half-cycle reference duration Perform a proportional conversion to obtain the actual resolution angle. The actual analytical angle Used to indicate waveform distortion points relative to the local zero-crossing reference time The angular position, expressed in degrees, is expressed by the following formula: .in, This refers to the angular range corresponding to the current half-cycle of the interaction.

[0109] Step S2300, the actual resolution angle Superimposed preset angle offset correction amount The corrected resolution angle is obtained. And according to the corrected analytical angle Determine the restoration control semantics within the target angle range specified in the preset angle range table. .

[0110] Specifically, this step aims to determine the actual resolution angle based on step S2200. The angular offset correction amount corresponding to the current power supply circuit. By performing combined operations, the corrected resolution angle is obtained. Based on the aforementioned corrected analytical angle Within the target angle range specified in the preset angle range table, determine the restored control semantics corresponding to the delay angle control content written by the master station into the grid voltage waveform. .

[0111] The restoration control semantics The control semantic items are pre-established in a preset angle interval table. Each control semantic item corresponds one-to-one with a target angle interval and is used to indicate the control actions that the air conditioning terminal should perform.

[0112] In practical implementation, in the scenario of energy-saving renovation of large public buildings, the main station writes delay angle control content to the air conditioning terminal through the power supply line. When this delay angle control content is transmitted on the grid voltage waveform, it is affected by the power supply loop length, cable impedance, and the connection status of the air conditioning inverter load, thus causing waveform distortion points. The corresponding angular position within the current half-cycle of the communication shifts relative to the angle set by the main station. If only the fixed angular boundary is applied to the actual analytical angle... Making a judgment can easily lead to the actual analytical angle being located near the boundary of the target angle interval. Falling into the adjacent target angle range leads to incorrect recognition of the air conditioning control action corresponding to the delay angle control content. Therefore, this step involves resolving the actual angle... Based on this, an angle offset correction amount corresponding to the current power supply circuit is introduced. For actual analytical angles After making the correction, the corrected resolution angle is then executed. Determine the reduction control semantics This is to improve the accuracy of restoring the content controlled by the delay angle. The actual analytical angle... Make corrections and apply them to the corrected analytical angle. Determining the target angle range and restoring control semantics The specific execution logic is as follows:

[0113] First, the controller retrieves a preset angle interval table. This table is used to establish the correspondence between the delay angle intervals within the current AC half-cycle and the control semantics. The delay angle intervals within the current AC half-cycle refer to the intervals corresponding to the current AC half-cycle. to Multiple target angle intervals are formed after the angle range is divided. This preset angle interval table will show the current AC half-cycle. to The angle range is divided into multiple non-overlapping target angle intervals, each corresponding to a control semantic. For example, the first target angle interval corresponds to the compressor start control semantic, the second target angle interval corresponds to the compressor stop control semantic, the third target angle interval corresponds to the four-way valve reversing control semantic, and the fourth target angle interval corresponds to the fan speed adjustment control semantic. Using the preset angle interval table, a correspondence is established between the delay angle control content written by the master station, the target angle intervals, and the air conditioning control actions.

[0114] Subsequently, the actual resolution angle is obtained. And obtain the angle offset correction amount corresponding to the current power supply circuit. The angle offset correction amount This is used to characterize the angular offset caused by cable impedance, distributed parameters, and the air conditioner load connection status in the current power supply circuit. Specifically, the master station continuously writes at least one calibration control frame with a known target angle to the air conditioner terminal within a preset calibration period. The air conditioner terminal calculates the actual resolved angle for each calibration control frame according to step S2200. And calculate each actual analytical angle. The angle difference between the current power supply circuit and the corresponding known target angle is calculated; subsequently, the angle difference is statistically processed to obtain the angle offset correction amount corresponding to the current power supply circuit. .

[0115] For example, when the main station continuously writes multiple target angles, all of which are... The calibration control frame corresponds to the actual resolved angle calculated by the air conditioning terminal. They are respectively , and When, the differences between the angles are respectively , and After averaging the angle differences, the angle offset correction amount corresponding to the current power supply circuit can be obtained. for .

[0116] In obtaining the actual resolution angle and angle offset correction amount Then, for the actual analytical angle After correction, the corrected resolution angle is obtained. The correction logic is as follows: The actual analytical angle is... Superimposed angle offset correction The corrected resolution angle is obtained. .

[0117] The above correction will reduce the angle offset introduced into the current power supply circuit. Compensation to actual resolution angle In the middle, make the corrected analytical angle It is closer to the target angle position corresponding to when the main station writes the grid voltage waveform.

[0118] Finally, based on the corrected resolution angle Determine the restoration control semantics corresponding to the target angle interval in the preset angle interval table. Wherein, the step of adjusting the analytical angle... The process of determining the target angle range refers to correcting the resolution angle. Compare the angles with the ranges of each target angle and correct the resolution angle. Includes the corrected resolution angle The target angle range. The restoration control semantics. It is not a free semantic generated on an ad-hoc basis, but rather a correspondence established in advance according to a preset angle interval table, derived from the corrected analytical angle. The control semantic items retrieved from the target angle interval.

[0119] Step S2400, according to the restoration control semantics In the local protocol library being called The corresponding control protocol item is retrieved from the database and combined with the local zero-crossing reference time. Generate air conditioning control interface signal This enables the controlled air conditioning terminal to perform the corresponding control action.

[0120] Specifically, this step will restore the control semantics from step S2300. This is converted into a drive pulse sequence that the controlled air conditioning terminal can receive, and then the local protocol library is invoked. The interface control parameters are used to generate air conditioning control interface signals for driving the compressor, fan, or four-way valve. This enables the controlled air conditioning terminal to execute and restore the control semantics. The corresponding control action is generated, and an electrical response corresponding to the control action is formed in the power supply circuit, so as to perform AI harmonic fingerprint recognition in the subsequent step S3000.

[0121] In practical implementation, at the terminal execution stage of air conditioning control in large public buildings, air conditioning terminals of different brands and with different interface types differ in their control input methods. To ensure the restoration of control semantics... This step, which maps to a specific controlled device, restores the control semantics. Converted into an air conditioning control interface signal that matches the target air conditioning terminal. And according to the local zero-crossing reference time output in step S2100 Determine the air conditioning control interface signal The start output time This ensures that the control execution action and subsequent AI harmonic fingerprint recognition correspond to the same control process. The specific execution logic is as follows:

[0122] First, obtain the restore control semantics. and in the local protocol library Searching for and restoring control semantics The control protocol entries corresponding to the currently controlled air conditioning terminal. The local protocol library. It pre-stores interface control parameters corresponding to different air conditioning terminals. These interface control parameters include at least the interface type, drive duration, drive sequence, and time interval between adjacent drives. For example, when the controlled air conditioning terminal uses relay contact control, the local protocol library... The local protocol library stores the relay closing sequence and duration corresponding to compressor start-up, compressor stop-up, four-way valve reversal, and fan speed adjustment; when the controlled air conditioning terminal uses infrared interface control, the local protocol library... The corresponding infrared code sequence and transmission duration are stored in the database.

[0123] After retrieving the restore control semantics Corresponding benchmark drive sequence Then, the local zero-crossing reference time output in step S2100 is called. The reference driving sequence For pre-stored in the local protocol library The control timing template in the middle is used to specify the current restore control semantics. The corresponding interface output content and output order. The local zero-crossing reference time. Used as a unified timing reference point within the current half-cycle of communication, for determining the air conditioning control interface signals. The output starting point. Then, according to the preset phase offset angle... and the half-cycle reference duration from step S1100 Determine the air conditioning control interface signal The start output time The starting output time The specific calculation formula is as follows: .

[0124] The preset phase offset angle Indicates air conditioning control interface signal relative to the local zero-crossing reference time The delayed output angle position is predetermined based on the access characteristics of the corresponding actuator of the controlled air conditioning terminal. Specifically, during the installation and commissioning phase, the same control action can be output at multiple candidate phase positions for the same controlled air conditioning terminal, and the corresponding starting current response can be collected; then, the phase position that makes the repeatability of the starting current response corresponding to the control action consistent can be selected as the phase offset angle corresponding to the controlled air conditioning terminal when executing the control action. The candidate phase position is set within the range of 20° to 160° of the current AC half-cycle to avoid the section where the voltage is close to zero near the zero crossing point and the section where the conduction period is shortened at the end of the half-cycle.

[0125] Determine the start output time Then, based on the benchmark driving sequence Generate air conditioning control interface signal When the reference drive sequence When multiple drive segments are included, each drive segment includes at least a corresponding output level, duration, and interval time between the previous drive segment. The multiple drive segments are arranged in a predetermined order; this predetermined order is determined by the local protocol library. Predefined to specify the output order of each driving segment, and not adjusted during execution. For example, when restoring control semantics... When the compressor starts, if the corresponding reference drive sequence Including a first drive segment, a second drive segment, and a third drive segment, the predetermined sequence can be defined as follows: first, output the first drive segment to activate the main control relay; then, after a predetermined interval, output the second drive segment to activate the compressor branch; finally, output the third drive segment to maintain the start-up state. For example, when the controlled air conditioning terminal uses infrared control, the predetermined sequence can be defined as the sequential output order of the preamble segment, function code field, and end segment. This step follows the predetermined sequence of each drive segment, starting from the initial output time. The process begins by sequentially outputting each driving segment, thereby forming and reconstructing the control semantics. Corresponding air conditioning control interface signal .

[0126] Step S3000: Acquire air conditioning control interface signal Controlled current sampling sequence after triggering and controlled voltage sampling sequence Extracting harmonic feature sets And construct harmonic fingerprint feature vectors The actual operating state matrix is ​​generated by combining the angle constraint range corresponding to the control action. .

[0127] Specifically, this step aims to monitor the harmonic characteristics generated by the controlled terminal in the physical execution state within the controlled loop. Utilizing the physical characteristic that the current waveform evolves with the action of the internal actuators, the harmonic characteristics of the air conditioning control interface signal in step S2400 are monitored. Controlled current sampling sequence after driving and controlled voltage sampling sequence Mapped to harmonic fingerprint features with operating condition identification By calculating the harmonic fingerprint features The correlation matching degree between the controlled terminal and the reference feature group determines the actual operating state matrix corresponding to the controlled terminal. This enables closed-loop monitoring of the operating status of the controlled terminal without the need for additional sensors.

[0128] Further, step S3000 includes:

[0129] Step S3100: Acquire air conditioning control interface signal Controlled current sampling sequence after triggering and controlled voltage sampling sequence Sampling sequence of controlled current Perform a Fast Fourier Transform to obtain the effective values ​​of each harmonic current. Combined with controlled voltage sampling sequence Extract the phase bias of each harmonic. Calculate the total harmonic distortion rate. Forming a harmonic feature set .

[0130] Specifically, this step aims to transfer the air conditioning control interface signal from step S2400 to the controlled branch circuit of the air conditioning system. Triggered controlled current sampling sequence and controlled voltage sampling sequence As a harmonic characteristic quantity, it is obtained by sampling the controlled current sequence. Perform a Fast Fourier Transform to obtain the effective values ​​of each harmonic current; and combine this with the controlled voltage sampling sequence. Extract the phase offset of each harmonic to form a harmonic feature set. And simultaneously extract the total harmonic distortion rate. The current change corresponding to the control action is converted into a harmonic characteristic input that can be used to identify the operating status of the air conditioner.

[0131] In the specific implementation process, in the feedback loop of the air conditioning control system for large public buildings, the end-effector receives signals from the air conditioning control interface. During the current analysis period after output, the controlled current sampling sequence of the air conditioning circuit is simultaneously acquired. With controlled voltage sampling sequence The controlled current sampling sequence is described above. The controlled voltage sampling sequence is used to characterize the current change of the corresponding load branch of the controlled air conditioning terminal during the execution of control actions. Used to characterize the change in supply voltage within the same time period, serving as a reference sequence for harmonic phase calculation.

[0132] After completing the synchronous acquisition, the controlled current sampling sequence is... A dynamic spectrum conversion is performed to obtain the spectral components of each harmonic current. The dynamic spectrum conversion is performed based on the starting output time in step S2400. Controlled current sampling sequence within the current analysis time period set for the time base A Fast Fourier Transform is performed to convert the current sample values ​​in the time domain into harmonic current spectral components in the frequency domain. The current analysis time period is the time during which the end-effector outputs the air conditioning control interface signal in step S2400. Then, a continuous sampling time interval is selected to obtain the electrical response corresponding to the control action.

[0133] The first is obtained through the Fast Fourier Transform. RMS value of subharmonic current .in, Represents the controlled current sampling sequence In the The effective value of the current at the second harmonic frequency is the first Amplitude characterization of the spectral components of the second harmonic current; Representing the harmonic order, it is the index of the spectral component of the harmonic current, with a value greater than or equal to 2 and less than or equal to the highest harmonic order. Positive integers. Indicates the highest harmonic order involved in spectrum calculation and harmonic feature construction.

[0134] Obtaining the effective values ​​of harmonic currents of each order Then, the total harmonic distortion rate is further calculated. The total harmonic distortion rate This is used to characterize the controlled current sampling sequence in the controlled air conditioning circuit during the current analysis period. The specific formula for calculating the overall distortion of the corresponding current waveform relative to the fundamental current component is as follows: The total harmonic distortion (THD) The calculation logic is as follows: First, calculate the second to the third... RMS value of subharmonic current The components are squared and summed separately, and then the sum is squared to obtain the combined effective value of the non-fundamental harmonic current components. Subsequently, the combined effective value of the non-fundamental harmonic current components is divided by the effective value of the fundamental current. This yields the ratio of the non-fundamental current component to the fundamental current component; finally, it is multiplied by 100%, and the ratio is converted into a percentage and output as the total harmonic distortion rate. .

[0135] The effective value of the fundamental current Yes, controlled current sampling sequence After performing a Fast Fourier Transform, at the power supply reference frequency The RMS current value extracted at the corresponding frequency point is used as a normalized reference value for each harmonic current component. The power supply reference frequency is also included. This indicates the reference frequency of the AC power supply currently connected to the controlled air conditioning circuit. In a power frequency supply scenario, Take the rated frequency of the power grid; the first The frequency of the second harmonic is The effective value of the fundamental current. The value is determined by the controlled current sampling sequence. As input, perform a Fast Fourier Transform, with the positioning frequency being the power supply reference frequency. The spectral components are calculated and converted into effective values ​​to obtain the effective value of the fundamental current. .

[0136] In calculating the total harmonic distortion rate Simultaneously, the controlled voltage sampling sequence As a phase reference sequence, the effective values ​​of each harmonic current component Extract the corresponding phase offset .in, Indicates the first The subharmonic current component relative to the controlled voltage sampling sequence The Middle The phase difference of the subharmonic voltage components. This phase bias. The specific extraction method is as follows: sample the controlled current sequence respectively. and controlled voltage sampling sequence Perform a Fast Fourier Transform to obtain the first... The phase angle of the second harmonic current spectral component and the first The phase angle of the second harmonic voltage spectrum component is calculated by subtracting the two. Phase offset of subharmonics.

[0137] Subsequently, the effective values ​​of each harmonic current were... With the corresponding phase offset Combining them to form a harmonic feature set Among them, the harmonic feature set It can be represented as This harmonic feature set This is used to characterize the harmonic response relationship corresponding to the conduction state of the controlled air conditioning circuit, the connection state of the actuator, and the change state of the electromagnetic load.

[0138] Step S3200, based on the harmonic feature set Total Harmonic Distortion Select a target set of harmonic orders, and then analyze the effective values ​​of the harmonic currents for each order within the target set of harmonic orders. The normalized harmonic amplitude is obtained by performing ratio conversion. and normalize the harmonic amplitude. Phase bias Total Harmonic Distortion Harmonic fingerprint feature vectors are constructed by combining elements in a preset order. .

[0139] Specifically, this step aims to process the harmonic feature set from step S3100. Total Harmonic Distortion By combining harmonic orders and corresponding control action time periods, a harmonic fingerprint feature vector is mapped to represent the operational state of the air conditioning actuators. This is to capture the differences in current response generated by the air conditioning circuit under control actions.

[0140] In the specific implementation process, during the state feedback phase of the air conditioning control system in large public buildings, the harmonic feature set is obtained. ,in, Then, based on the preset tiered filtering rules, from the 1st to the 2nd... The target harmonic order set for fingerprint construction is selected from the subharmonics. The preset order selection rule is as follows: based on the effective values ​​of the corresponding harmonic currents of the compressor, four-way valve, or fan in historical control samples. and phase bias The amplitude of the change is used to select the harmonic order whose amplitude exceeds a preset threshold as the target harmonic order. This method establishes a correspondence between the target harmonic order and the actions of the air conditioning actuators.

[0141] To reduce the impact of power supply amplitude fluctuations on harmonic fingerprint feature vectors Due to the influence of the structure, this step uses the effective value of the fundamental current in step S3100. As a normalization benchmark, the effective values ​​of harmonic currents of each order in the target harmonic order set are... By performing ratio conversion, the normalized harmonic amplitude is obtained. The normalized harmonic amplitude Indicates the first RMS value of subharmonic current Relative to the effective value of the fundamental current The amplitude ratio. For the first harmonic order selected into the target harmonic order set... The first harmonic, whose corresponding normalized harmonic amplitude is denoted as By using the above ratio conversion, a unified comparison benchmark is achieved for the harmonic amplitudes generated by different air conditioning terminals under different current bases.

[0142] After completing the ratio conversion, this step will convert the normalized harmonic amplitude values ​​corresponding to each target harmonic order set. Phase bias and total harmonic distortion Harmonic fingerprint feature vectors are constructed by combining elements in a preset order. The harmonic fingerprint feature vector Specifically, it is expressed as follows: .in, Indicates the target harmonic order involved in the construction; This indicates the number of target harmonic orders in the target harmonic order set; Indicates the first Normalized harmonic amplitude of the subharmonic; Indicates the first Phase offset of the subharmonic; superscript This indicates transpose, used to construct the above features into a column vector.

[0143] The preset arrangement order includes the harmonic order order and the corresponding time order of the control actions. The harmonic order order is used to ensure that the harmonic fingerprint feature vectors generated by different controlled air conditioning terminals are consistent. They have a consistent dimensional position, and their setting rule is: arranged according to harmonic order from smallest to largest, and the normalized harmonic amplitude of the same harmonic order is used. With phase bias Adjacent settings; the timing sequence of the control actions is used to ensure the harmonic fingerprint feature vector The air conditioning control interface signal in step S2400 The start output time One-to-one correspondence, the setting rule is: based on the starting output time Using a time reference, harmonic features extracted within the corresponding current analysis time period are numbered and arranged according to the chronological order of each control action. By simultaneously introducing the harmonic order sequence and the corresponding time sequence of control actions, the air conditioning control interface signal can be... The output electrical response is fixed in a uniform characteristic coordinate system.

[0144] Step S3300, based on the harmonic fingerprint feature vector and controlled voltage sampling sequence Determine the angle constraint interval corresponding to the control action, and then use the harmonic fingerprint feature vector. The harmonic characteristic components located within the angle constraint interval are combined into a joint feature, and the actual operating state matrix is ​​generated according to the state mapping rule. .

[0145] Specifically, this step aims to extract the harmonic fingerprint feature vector from step S3200 from the controlled branch circuit of the air conditioner. The controlled voltage sampling sequence from step S3100 will be used as the state determination input. As a basis for angle positioning, the normalized harmonic amplitude corresponding to each target harmonic order is... Phase bias and total harmonic distortion Correspondence analysis is performed to obtain the actual operating state matrix characterizing the action results of the air conditioning actuators. To determine the air conditioning control interface signal output in step S2400. Has the corresponding control action actually been completed?

[0146] In the specific implementation process, in the scenario of energy-saving retrofitting of existing circuits in air conditioning systems of large public buildings, the control device completes the air conditioning control interface signal in step S2400. After the interface is driven, the compressor, four-way valve, fan motor, and their corresponding drive branches inside the air conditioning terminal will generate harmonic characteristic changes corresponding to the control action in the controlled voltage and controlled current. Because the mechanical action processes corresponding to different actuators are different, even if current exists in all circuits, the normalized harmonic amplitude of the corresponding target harmonic order will vary. Phase bias and total harmonic distortion Differences still exist. Therefore, this step does not use whether the circuit is conductive as the sole criterion, but instead uses the harmonic fingerprint characteristics formed after delay angle control as the basis for state determination. It identifies states where control commands have been output but the executing components have not completed the corresponding actions, in order to generate the actual operating state matrix. The actual operating state matrix The specific generation logic is as follows:

[0147] First, obtain the harmonic fingerprint feature vector. The harmonic fingerprint feature vector This includes the harmonic characteristic components corresponding to each target harmonic order; the harmonic characteristic components include normalized harmonic amplitudes. Phase bias and total harmonic distortion Simultaneously, the controlled voltage sampling sequence corresponding to the same control action is acquired. .

[0148] Subsequently, the local zero-crossing reference time output in step S2100 is used. As a unified timing reference point within the current AC half-cycle, and in conjunction with the controlled voltage sampling sequence The time interval between adjacent zero-crossing points is used to construct the half-cycle angle reference for the current analysis cycle. Each pair of adjacent local zero-crossing reference times... Each half-cycle corresponds to an alternating current cycle, and the duration of this half-cycle is converted into a half-cycle angle interval from 0° to 180°. The first and second half-cycles are converted independently, forming multiple consecutive half-cycle angle intervals. Based on this, a harmonic fingerprint feature vector is constructed. The harmonic characteristic components corresponding to each target harmonic order are converted to their corresponding half-cycle angle positions according to the half-cycle angle interval in which their sampling time falls. In this way, the harmonic fingerprint features extracted in different control cycles are all angularly located relative to the same zero-crossing starting point, thereby reducing the harmonic position offset caused by changes in power supply cycle, offset of sampling start time, and differences in the timing of actuator actions, and enabling the harmonic fingerprint results corresponding to each control action to be compared under a unified half-cycle angle reference.

[0149] After unifying the half-cycle angle reference, an operation state mapping process based on angle constraints is performed. This operation state mapping process refers to first mapping the operation state according to the controlled voltage sampling sequence. Generate the angle constraint interval corresponding to the current control action, and then generate the harmonic fingerprint feature vector. The harmonic characteristic components corresponding to each target harmonic order are filtered, repositioned, and combined according to the angle constraint interval to form a joint feature for state determination. Finally, the actual operating state matrix is ​​output according to the pre-established state mapping rules. .

[0150] The angle constraint range is represented by the controlled voltage sampling sequence. The determined semi-circle start point, semi-circle length, and angular positioning results for each target observation interval. The angular constraint interval is constructed using the local zero-crossing reference time. As the starting point of the current half-cycle, the controlled voltage sampling sequence is then used. Two adjacent local zero-crossing reference times The time interval between intervals determines the current half-cycle duration, which is then converted into a half-cycle angle range from 0° to 180°. After forming multiple consecutive half-cycle angle ranges, the air conditioning control interface signal is used as the basis for the calculation. The output start point and duration are used to determine the target observation window corresponding to the current control action; then, the target observation window is mapped to the start and end angle positions in the corresponding half-cycle angle interval to obtain the angle constraint interval corresponding to the current control action. The angle constraint interval is used to limit the effective positions of the harmonic feature components corresponding to each target harmonic order in the state determination, so that the same harmonic feature component only participates in identification within the half-cycle angle range corresponding to the control action, thereby making the harmonic fingerprint feature... The air conditioning control interface signal formed in step S2400 The actual driving actions maintain a corresponding relationship.

[0151] The joint feature represents the normalized harmonic amplitude corresponding to each target harmonic order extracted within the angle constraint interval. Phase bias and total harmonic distortion The combined results of changes in the adjacent analysis periods.

[0152] The state mapping rule is as follows: For each air conditioner action, a reference feature set corresponding to the action in the normal state is pre-established, and the joint feature to be determined is compared with the reference feature set to determine the actual operating state matrix. The state results are as follows. The reference feature set can be obtained by calibrating the historical sampling results of similar air conditioning terminals during compressor start-up and shutdown, compressor loading, four-way valve reversal, and fan speed switching. The comparison method can use the difference value calculation method. For example, in the four-way valve reversal action, the reversal-sensitive harmonic order can be selected as the target harmonic order set. When the normalized harmonic amplitude of the current joint feature is within the target observation window... The difference between the value and the reference feature group after commutation is not greater than the first determination threshold, and the phase offset is... When the direction of change is consistent with the direction of the reversal command, the actual running state matrix will be... The state of the four-way valve in the circuit is recorded as commutation complete; when the circuit is open but the difference value is greater than the first determination threshold, or the phase offset is... If no corresponding directional change occurs, the four-way valve status is recorded as a reversing abnormality.

[0153] The actual operating state matrix This represents the actual execution result of the controlled air conditioning terminal within the current control cycle. This actual operating state matrix... It includes at least one or more of the following: compressor loading state, four-way valve reversing state, indoor fan rotation state, and anomaly marking state. Specifically, the compressor loading state indicates whether the compressor has completed the corresponding start-stop or loading action; the four-way valve reversing state indicates whether the four-way valve has completed the target direction switch; the indoor fan rotation state indicates whether the fan has entered the corresponding rotation state according to the target gear; and the anomaly marking state indicates whether the combined feature obtained after the current control action is consistent with the corresponding reference feature group.

[0154] In this way, after completing the delay angle control, the control device can collect harmonic characteristics corresponding to the control action on the existing power supply line, and then use the harmonic fingerprint characteristics... and controlled voltage sampling sequence The actual operating status of the air conditioning terminal is determined, thereby enabling the identification of the air conditioning terminal's operating status.

[0155] Step S4000, according to the initial setting command and actual operating state matrix Calculate the consistency confidence score Based on consistency confidence Harmonic feature set Obtain the exception type flag And according to the exception type flag Generate remedial control flow Provide feedback.

[0156] Specifically, this step aims to process the initial setting instructions from step S1200, which characterize the user's expected goal. Compared with the actual operating state matrix characterizing the microscopic physical conditions of the equipment from step S3300 As a comparison object, the state deviation between the two is converted into state harmonic potential energy using difference calculation. Mapping invisible execution deviations to consistency confidence levels Evolutionary gradient vector And when an anomaly type flag is identified It is then decoded to contain remedial control instructions. Remedial control flow At the source of the control closed loop, the expected target and the actual operating condition are verified and aligned, providing the control and regulation system with closed-loop self-healing regulation input.

[0157] Further, step S4000 includes:

[0158] Step S4100: Obtain initial setting command and actual operating state matrix , initial setting command Encode as target state vector From the actual operating state matrix Extract the actual state vector And based on the target state vector and actual state vector Calculate state harmonic potential energy To obtain the consistency confidence level .

[0159] Specifically, this step aims to perform a unified correlation analysis between the control settings of the air conditioning system in a large public building and the actual operating status obtained by AI harmonic fingerprint recognition, and to integrate the initial setting instructions from step S1200. With the actual operating state matrix from step S3300 The state vector is converted into a unified encoding, and the state deviation between the two, i.e., the state harmonic potential energy, is calculated. This approach maps abstract instruction execution deviations to measurable potential energy displacements, aiming to eliminate state blind spots caused by missing physical links and provide physically deterministic consistency confidence for the closed-loop self-healing of air conditioning systems. enter.

[0160] In the specific implementation process, during the closed-loop self-verification phase of the air conditioning system, this step needs to determine the signal from the air conditioning control interface in step S2400. Does the corresponding target state match the actual state output by AI harmonic fingerprint recognition in step S3300? The specific processing procedure is as follows:

[0161] First, obtain the initial setting command. and the actual operating state matrix The initial setting command, in particular... The target control command used to represent the current air conditioning equipment includes at least one of the following: target air conditioning operating mode, target temperature range, target fan speed, target compressor start / stop status, and target four-way valve status; the actual operating status matrix Step S3300 is the harmonic fingerprint feature vector The status recognition results obtained after performing AI harmonic fingerprint recognition include compressor operating status, four-way valve reversing status, indoor fan operating status, and abnormal marking status.

[0162] To enable the initial setting command With the actual operating state matrix With a unified comparison benchmark, this step involves the initial setting instructions. The target state vector is obtained by performing instruction encoding processing. ; for the actual operating state matrix Perform state extraction processing to obtain the actual state vector. Wherein, the target state vector The actual state vector is used to represent the target control state that the air conditioning terminal is required to achieve in step S1200. This is used to indicate the actual operating status identified in step S3300.

[0163] The instruction encoding process refers to encoding the initial setting instructions according to a unified state encoding order. The target operating mode, target compressor status, target four-way valve status, and target fan status are mapped to corresponding status code values ​​and arranged in a fixed order to form a target state vector. The state extraction process refers to the extraction of data from the actual operating state matrix. Extract the recognition results corresponding to the state coding sequence, convert the compressor state, four-way valve state, indoor fan state, and operating mode state into corresponding state coding values, and then sort them according to the target state vector. The same order forms the actual state vector. .

[0164] The target state vector With the actual state vector A consistent status coding sequence is used. This status coding sequence includes at least the compressor status code, four-way valve status code, fan status code, and operating mode code. By using a consistent status coding sequence, the control command side and the status identification side can be unified into the same status coordinate system. Specifically, the compressor status code is used to indicate whether the compressor is in a stopped or loaded state; the four-way valve status code is used to indicate whether the four-way valve is in a cooling or heating position; the fan status code is used to indicate whether the fan is in a stopped state or rotating at the corresponding speed; and the operating mode code is used to indicate whether the air conditioner is in cooling mode, heating mode, fan-only mode, or standby mode.

[0165] Furthermore, to reflect the air conditioning control interface signal The start output time The impact on state determination, this step involves constructing the actual state vector. At the same time, introduce time period markers corresponding to control actions. The control action corresponds to a time period marker. Based on the air conditioning control interface signal The start output time And the sampling cutoff time corresponding to the current state identification in step S3300. Confirmed. By introducing time period markers corresponding to control actions. This can limit the state changes after the delay angle control signal is output to the observation time window of the current control action, thus avoiding the inclusion of the residual state formed by the previous control action into the current consistency determination process.

[0166] After completing the target state vector With the actual state vector After construction, this step performs a difference calculation on the two to obtain the state harmonic potential energy. The state harmonic potential energy Used to characterize the target state vector With the actual state vector The degree of mismatch between them is calculated using the following formula: Among them, superscript Indicates transpose; This represents the state weight matrix, used to characterize the order of influence of different state terms on the air conditioning control results.

[0167] The state weight matrix The weights are set according to the order of influence of compressor status, four-way valve status, fan status, and operating mode status on the air conditioning control results. The state item corresponding to compressor status is assigned the first weight; the state item corresponding to four-way valve status is assigned the second weight; the state item corresponding to fan status is assigned the third weight; and the state item corresponding to operating mode status is assigned the fourth weight. When the target state vector... With the actual state vector When the encoded values ​​of a certain state item are inconsistent, the difference between the encoded values ​​of that state item in the two vectors is expressed through the state weight matrix. Weighted and including state harmonic potential energy The calculation results.

[0168] Wherein, the first weight, second weight, third weight, and fourth weight are normalized weight parameters, with values ​​ranging from 0 to 1, and satisfying the following conditions: the first weight is greater than the second weight, the second weight is greater than the third weight, the third weight is greater than the fourth weight, and the sum of all weights is 1. For example, the first weight can be 0.4, the second weight can be 0.3, the third weight can be 0.2, and the fourth weight can be 0.1. Through the above settings, key execution components affecting the cooling / heating switching results and energy consumption results can occupy corresponding proportions in the consistency determination.

[0169] Obtaining state harmonic potential energy Then, this step converts it into a consistency confidence level. This consistency confidence level The calculation method is: 1 divided by 1 and the state harmonic potential energy. The sum of the states. Therefore, when the state harmonic potential energy... When the value is small, the consistency confidence level Larger and approaching 1; conversely, when the state harmonic potential energy When the value is large, the consistency confidence level It is relatively small and approaches 0, thus realizing the transformation of the difference between the control target and the actual state into a unified value.

[0170] When consistency confidence Not less than the preset consistency threshold At that time, determine the air conditioning control interface signal output in step S2400. The air conditioning terminal has reached the target control state; when the consistency confidence level is... Less than the preset consistency threshold When the current control action is not completed, or the air conditioning terminal is in at least one of the following inconsistent states: the compressor is not actually running, the four-way valve has not completed its state switching, or the fan has not reached the target operating state, the determination result is output to the subsequent abnormal handling steps. The preset consistency threshold is mentioned above. The preset consistency threshold is determined based on the statistical distribution of consistency confidence levels in historical control samples and normal operation samples collected during system debugging, and its value ranges from 0 to 1. Preferably, the preset consistency threshold... Values ​​between 0.75 and 0.95 are acceptable.

[0171] Step S4200, based on the consistency confidence level Solving the evolution gradient vector , evolve gradient vector Harmonic feature set Perform joint discrimination to obtain anomaly type flags. .

[0172] Specifically, this step aims to assess the consistency confidence level derived from step S4100, which characterizes the degree of matching between the expected target and the actual response in the air conditioning control closed loop. The harmonic response set at the physical level from the execution loop in step S3100 Perform association analysis to determine the consistency confidence level in the time domain. The change process is mapped to an anomaly type flag. The corresponding physical failure types are identified to distinguish between abnormal control signal transmission and abnormal actuator operation, providing a basis for handling abnormalities in the air conditioning control circuit.

[0173] In the specific implementation process, when the consistency confidence level Less than the preset consistency threshold When an anomaly is detected in the air conditioning control loop corresponding to the current delay angle control, anomaly physical attribute identification processing is performed. This anomaly physical attribute identification processing includes two steps: consistency confidence change calculation and harmonic feature joint discrimination.

[0174] The specific logic for calculating the change in consistency confidence is as follows: First, the consistency confidence... The evolution gradient vector is obtained by calculating the changes within the current observation period. The evolution gradient vector Used to characterize consistency confidence The formulas for calculating the direction and magnitude of change over time are as follows: in, This represents the change in the consistency confidence level over the observation period. Indicates the corresponding time interval. This represents the evolution correction coefficient. The evolution correction coefficient... The power grid noise intensity is determined based on the grid noise intensity within the current observation period and is a real number greater than 0. Preferably, its value ranges from 0.5 to 1. The grid noise intensity is calculated from the collected power supply waveform data, which includes at least one of the following: voltage fluctuation, current fluctuation, and fundamental zero-crossing offset. This evolution correction coefficient... The specific value selection logic is as follows: the collected noise amplitude is compared with the preset reference noise amplitude to obtain the evolution correction coefficient Θ. The preset reference noise amplitude is determined based on normal power supply waveform samples collected during the system commissioning phase; specifically, it is obtained by statistically analyzing the noise amplitude over multiple normal operating cycles. The evolution correction coefficient is then introduced. This allows the evolution gradient vector to be... It reflects the consistent trend of changes corresponding to the current control action, and does not misjudge random fluctuations caused by background noise as abnormal changes.

[0175] Obtaining the evolution gradient vector Next, this step performs joint harmonic feature discrimination, and its specific execution logic is as follows: First, the evolved gradient vector is... With harmonic feature set Perform time alignment. Time alignment refers to aligning the consistency confidence level... Corresponding observation time period and harmonic feature set By matching the corresponding sampling time periods, the consistency change is correlated with the harmonic response caused by the same control action.

[0176] When the evolution gradient vector The modulus value is greater than the preset mutation threshold. And the harmonic feature set corresponding to this observation period When none of the middle harmonic components meet the effective identification criteria, the processor determines that the current anomaly originates from the delay angle control signal not entering the air conditioning execution loop, and sets the anomaly type flag accordingly. This is marked as a control signal transmission anomaly. The failure of each harmonic component to meet the effective identification criteria means that the harmonic amplitude of the corresponding harmonic component is lower than the preset identification threshold, or that a harmonic characteristic distribution corresponding to the current control action has not been formed within multiple consecutive sampling periods. This type of anomaly corresponds to situations where there is link interference, injection failure, or interface non-conductivity during the delay angle control output process.

[0177] When the evolution gradient vector The modulus value is not greater than the preset mutation threshold. And the harmonic feature set corresponding to this observation period The processor determines that the current anomaly originates from the failure of the air conditioning actuator to complete its operation, and marks the anomaly type as an error. The processor identifies the anomaly as originating from the failure of the air conditioning actuator to complete its operation. However, if the harmonic response deviates from the preset reference harmonic characteristics in terms of harmonic amplitude deviation, phase shift, or inter-order amplitude ratio at the corresponding order, the processor determines that the anomaly is due to the air conditioning actuator not completing its operation. The operation is marked as abnormal. Abnormal operation includes at least one of the following: compressor not actually starts, four-way valve not completing reversal, and fan not reaching the target speed. This type of abnormality corresponds to the situation where the delay angle control signal has entered the air conditioning execution circuit, but the controlled component has not completed the operation according to the target state. The preset reference harmonic characteristics are established based on normal control samples collected during the system debugging phase. These normal control samples include at least harmonic characteristic data corresponding to compressor start-up, four-way valve reversal, and fan speed switching. The preset deviation range is determined based on the statistical distribution of the amplitude, phase, and amplitude ratio between orders of each harmonic under the corresponding control action. Specifically, interval statistics are performed on the harmonic parameters corresponding to the same control action in the normal control samples to obtain the allowable deviation range for each harmonic parameter, and this allowable deviation range is used as the preset deviation range.

[0178] Step S4300, based on the exception type flag The delay angle control parameters are reorganized to generate remedial control commands. and will provide remedial control instructions Encapsulated as a remedial control flow Provide feedback.

[0179] Specifically, this step aims to utilize the exception type flag from step S4200. Identify the source of the anomaly in the current air conditioning control loop, and generate remedial control commands by combining the phase deviation retained from the previous control cycle and the harmonic parameter deviation used to determine the abnormal operation of the actuators. and by remedial control flow The remedial control command Feedback is sent to the input of the control system to correct the delay angle control parameters, thereby completing the end-of-loop repair of the control loop.

[0180] In the specific implementation process, this step is based on the exception type flag. Perform parameter reconfiguration processing. The parameter reconfiguration processing includes physical deviation trace extraction, remedial control command generation, and remedial control flow encapsulation and feedback.

[0181] The specific logic for extracting physical deviation traces is as follows: First, obtain the anomaly type flag. When the exception type flag is displayed. In order to control signal transmission abnormalities, the processor retrieves the actual resolution angle in step S2200. And combined with the target delay angle corresponding to the current control task Perform the difference operation to obtain the residual phase deviation. The residual phase deviation This is used to characterize the phase offset of the control signal formed in the transmission path and serves as the input for subsequent delay angle compensation. The target delay angle... The target action type is determined based on the current air conditioning control task. Specifically, the controller retrieves the delay angle setting value corresponding to the target action from the control task parameters according to a preset control mapping rule. The control task parameters include at least one of the following: operating mode, target temperature, fan speed, compressor start / stop status, and four-way valve reversing status.

[0182] The specific logic for generating the remedial control command is as follows: based on the residual phase deviation... The target delay angle for the next control cycle is compensated to obtain the optimized delay angle. and the optimized delay angle Write the angle parameter field corresponding to the delay angle control command to generate the remedial control command. The remedial control command Optimization of delay angle Determined according to the following formula: Among them, optimizing the delay angle This indicates the optimized delay angle of the output to the modulation circuit. This represents the compensation coefficient. The compensation coefficient... Determined based on power grid frequency fluctuation information, used to limit the optimized delay angle. Relative to target delay angle The correction amount enables the remedial control command. This compensation coefficient corresponds to the current power supply frequency fluctuation state. The specific execution logic is as follows: extract the frequency offset within the current observation period, compare it with the preset reference frequency fluctuation value, and determine the compensation coefficient based on the comparison result. The value of the preset reference frequency fluctuation is determined based on the normal power supply frequency data collected during the system debugging phase. Specifically, when the air conditioning execution circuit is in normal operation, the power grid frequency data is continuously collected within multiple control cycles, the frequency offset corresponding to each control cycle is calculated, and the frequency offset is statistically analyzed to obtain the preset reference frequency fluctuation value.

[0183] When the exception type flag When the abnormality is characterized by the operation of the actuator, the remedial control command is applied according to the harmonic amplitude deviation, phase deviation, or inter-order amplitude ratio deviation used to determine the abnormality in step S4200. Additional corrections are performed within the delay angle conduction interval to enable remedial control commands. The corresponding conduction duration is increased within a preset range to ensure that the physical response of the actuator meets the preset action judgment conditions. The additional correction refers to shifting the start or end point of the delay angle within the same control cycle without changing the target control category, thereby altering the equivalent action of the current control signal. The preset range is determined based on normal action samples and insufficient action samples collected during the system debugging phase. Specifically, it involves statistically analyzing the conduction duration intervals when the target action is completed and when the target action is not completed under the same control action, and taking the delay angle correction interval corresponding to the action compensation between the two as the preset range.

[0184] Generate remedial control instructions Next, this step performs remedial control flow encapsulation and feedback, the specific execution logic of which is as follows: the remedial control command is... Encapsulated as a remedial control flow And feed back to the sending entry corresponding to step S1000, so that the remedial control flow... This triggers a new round of delay angle encoding and transmission. The encapsulation refers to the process of encapsulating the remedial control command... Write the data into the recovery control flow according to the preset data format. The preset data format includes at least an action category identifier, an optimized delay angle parameter, a control cycle identifier, and a verification field. Through this method, the control system converts the anomaly identification result into a delay angle correction for the next control cycle and resends the corrected delay angle control command into the transmission link, ensuring that control signal transmission anomalies and actuator action anomalies correspond to different remedial paths.

[0185] Example 2:

[0186] This embodiment, based on Embodiment 1, provides an air conditioning control system based on delay angle control and AI harmonic fingerprint recognition, such as... Figure 2 As shown, the system includes a carrier voltage modulation module, an instruction parsing module, a status monitoring module, and an anomaly diagnosis and compensation module.

[0187] The carrier voltage modulation module is used to adjust the real-time frequency of the power grid. Determine the reference time for voltage zero crossing Combined with initial setting instructions and remedial control instructions Generate theoretical delay angle According to the theoretical delay angle Get the trigger time offset This drives the power switching devices to conduct, forming a carrier voltage waveform. .

[0188] The instruction parsing module is used to analyze the carrier voltage waveform. Determine the local zero-crossing reference time and waveform distortion points To solve for the actual analytical angle Based on actual analytical angle Determine the reduction control semantics And based on the restoration control semantics Generate air conditioning control interface signal

[0189] The status monitoring module is used to collect signals from the air conditioning control interface. Controlled current sampling sequence after triggering and controlled voltage sampling sequence Extracting harmonic feature sets And construct harmonic fingerprint feature vectors The actual operating state matrix is ​​generated by combining the angle constraint range corresponding to the control action. .

[0190] The anomaly diagnosis and compensation module is used to perform initial setting instructions. and actual operating state matrix Calculate the consistency confidence score Based on consistency confidence Harmonic feature set Obtain the exception type flag And according to the exception type flag Generate remedial control flow Provide feedback.

[0191] The parts of the technical solutions provided in the embodiments of this application that are consistent with the implementation principles of corresponding technical solutions in the prior art have not been described in detail to avoid excessive elaboration.

[0192] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the invention. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An air conditioner control method based on delay angle control and AI harmonic fingerprinting, characterized by, include: The voltage zero-crossing reference time is determined based on the real-time frequency of the power grid. The theoretical delay angle is generated by combining the initial setting command and the remedial control command. The trigger time offset is obtained based on the theoretical delay angle, and the power switching device is driven to conduct to form a carrier voltage waveform. The local zero-crossing reference time and waveform distortion point are determined based on the carrier voltage waveform to calculate the actual resolution angle. The restored control semantics are determined based on the actual resolution angle, and the air conditioning control interface signal is generated based on the restored control semantics. Collect the controlled current sampling sequence and controlled voltage sampling sequence after the air conditioner control interface signal is triggered, extract the harmonic feature set and construct the harmonic fingerprint feature vector, and generate the actual operating state matrix by combining the angle constraint interval corresponding to the control action; The consistency confidence level is calculated based on the initial setting instructions and the actual operating state matrix. An anomaly type flag is obtained based on the consistency confidence level and harmonic feature set. A remedial control flow is then generated based on the anomaly type flag for feedback.

2. The air conditioning control method based on delay angle control and AI harmonic fingerprint recognition according to claim 1, characterized in that, The method for forming the carrier voltage waveform includes: The half-cycle search window is determined based on the real-time frequency of the power grid. Curvature discrimination processing is performed on the collected voltage sampling sequence to determine the zero-crossing candidate time. The voltage zero-crossing reference time is generated by combining the time delay compensation of the sampling link. A composite control quantity is generated based on the initial setting command and the remedial control command, and the composite control quantity is mapped to the theoretical delay angle through a preset mapping function; The power grid cycle duration is determined based on the real-time frequency of the power grid, and the trigger time offset is obtained by combining the theoretical delay angle and the link compensation time. Timing begins based on the voltage zero-crossing reference moment. When the timing reaches the trigger time offset, the power switching device is driven to conduct, forming a carrier voltage waveform.

3. The air conditioning control method based on delay angle control and AI harmonic fingerprint recognition according to claim 2, characterized in that, The method for curvature discrimination processing includes: The half-cycle reference duration is determined based on the real-time frequency of the power grid, and a half-cycle search window is constructed based on the half-cycle reference duration. Within the half-cycle search window, sampling points with voltage values ​​close to zero are selected from the voltage sampling sequence as candidate points; the second-order difference is calculated based on the relationship between the changes of adjacent sampling values ​​before and after each candidate point to determine the zero-crossing candidate time.

4. The air conditioning control method based on delay angle control and AI harmonic fingerprint recognition according to claim 1, characterized in that, The method for generating the air conditioning control interface signal includes: A voltage sampling sequence is formed based on the carrier voltage waveform. The local zero-crossing reference time is determined by polarity determination, and the waveform distortion points are extracted based on the second-order difference of the voltage sampling sequence. Calculate the time difference between the waveform distortion point and the local zero-crossing reference time, and convert the time difference and half-cycle reference duration to obtain the actual resolution angle; The actual resolution angle is superimposed with a preset angle offset correction amount to obtain the corrected resolution angle, and the restoration control semantics are determined according to the target angle interval to which the corrected resolution angle belongs in the preset angle interval table. Based on the restored control semantics, the corresponding control protocol item is retrieved from the local protocol library and combined with the local zero-crossing reference time to generate the air conditioning control interface signal, so that the controlled air conditioning terminal can perform the corresponding control action.

5. The air conditioning control method based on delay angle control and AI harmonic fingerprint recognition according to claim 4, characterized in that, The method for determining polarity includes: Based on the carrier voltage waveform, a voltage sampling sequence arranged in chronological order is formed; for two adjacent sample values ​​in the voltage sampling sequence, it is determined whether their signs have changed; When two adjacent sampled values ​​are one positive and one negative, or the previous sampled value is zero and the next sampled value is non-zero, or the previous sampled value is non-zero and the next sampled value is zero, it is determined that there is a zero-crossing position in the adjacent sampling interval. The zero-crossing position is determined as the local zero-crossing reference time.

6. The air conditioning control method based on delay angle control and AI harmonic fingerprint recognition according to claim 1, characterized in that, The method for generating the actual operating state matrix includes: The controlled current sampling sequence and controlled voltage sampling sequence are collected after the air conditioner control interface signal is triggered. A fast Fourier transform is performed on the controlled current sampling sequence to obtain the effective value of each harmonic current. The phase offset of each harmonic is extracted by combining the controlled voltage sampling sequence, and the total harmonic distortion rate is calculated to form a harmonic feature set. The target harmonic order set is selected based on the harmonic feature set and the total harmonic distortion rate. The effective values ​​of each harmonic current in the target harmonic order set are converted by ratio to obtain the normalized harmonic amplitude. The normalized harmonic amplitude, phase offset and total harmonic distortion rate are combined in a preset order to construct the harmonic fingerprint feature vector. The angle constraint interval corresponding to the control action is determined based on the harmonic fingerprint feature vector and the controlled voltage sampling sequence. The harmonic feature components located within the angle constraint interval in the harmonic fingerprint feature vector are combined into a joint feature, and the actual operating state matrix is ​​generated according to the state mapping rule.

7. The air conditioning control method based on delay angle control and AI harmonic fingerprint recognition according to claim 6, characterized in that, The method for calculating the total harmonic distortion rate includes: Collect the controlled current sampling sequence after the air conditioner control interface signal is triggered; perform a fast Fourier transform on the controlled current sampling sequence to obtain the effective value of the fundamental current and the effective values ​​of the second to Nth harmonic currents; The fundamental current RMS value is the RMS current value extracted at the frequency point corresponding to the power supply reference frequency; the power supply reference frequency is the reference frequency of the AC power supply connected to the current controlled air conditioning circuit. The effective values ​​of the 2nd to Nth harmonic currents are squared and summed, and then the sum is squared to obtain the composite effective value. Divide the synthesized effective value by the effective value of the fundamental current and multiply by 100% to obtain the total harmonic distortion rate.

8. The air conditioning control method based on delay angle control and AI harmonic fingerprint recognition according to claim 1, characterized in that, The method for generating the remedial control flow includes: The initial setting command and the actual running state matrix are obtained. The initial setting command is encoded into a target state vector. The actual state vector is extracted from the actual running state matrix. The state harmonic potential energy is calculated based on the target state vector and the actual state vector to obtain the consistency confidence. The evolutionary gradient vector is calculated based on the consistency confidence level. The evolutionary gradient vector and harmonic feature set are jointly discriminated to obtain the anomaly type indicator. Based on the anomaly type flag, the delay angle control parameters are reorganized to generate a remedial control command, which is then encapsulated into a remedial control flow for feedback.

9. The air conditioning control method based on delay angle control and AI harmonic fingerprint recognition according to claim 8, characterized in that, The joint discrimination method includes: When the consistency confidence is less than the preset consistency threshold, the evolution gradient vector is obtained based on the change in consistency confidence during the current observation period, the corresponding time interval, and the evolution correction coefficient. The observation time period corresponding to the consistency confidence is matched with the sampling time period corresponding to the harmonic feature set. When the magnitude of the evolution gradient vector is greater than the preset mutation threshold and the harmonic components of each order in the harmonic feature set do not meet the effective identification conditions, the abnormality type flag is determined as a control signal transmission abnormality. If the magnitude of the evolution gradient vector is not greater than a preset mutation threshold, and there are harmonic components in the harmonic feature set that correspond to the target control action, but the amplitude deviation, phase shift, or inter-order amplitude ratio of the harmonic components relative to the preset reference harmonic features exceeds a preset deviation range, the abnormality type flag will be determined as an abnormal action of the execution component.

10. The air conditioning control method based on delay angle control and AI harmonic fingerprint recognition according to claim 8, characterized in that, The method for parameter recombination includes: When the anomaly type is marked as a control signal transmission anomaly, the actual resolution angle is retrieved, and the difference between the actual resolution angle and the target delay angle corresponding to the current control task is calculated to obtain the residual phase deviation. Based on the residual phase deviation, the target delay angle of the next control cycle is compensated to obtain the optimized delay angle, and the optimized delay angle is written into the angle parameter field of the delay angle control command to generate a remedial control command. When the anomaly type flag indicates an abnormal action of the executing component, additional corrections are performed on the delay angle conduction interval in the remedial control command to adjust the start or end point of the delay angle and generate a remedial control command. The remedial control command is encapsulated into a remedial control flow according to a preset data format and fed back to the sending entry.

11. An air conditioning control system based on delay angle control and AI harmonic fingerprint recognition, used to implement the air conditioning control method based on delay angle control and AI harmonic fingerprint recognition as described in any one of claims 1-10, characterized in that, The system includes a carrier voltage modulation module, an instruction parsing module, a status monitoring module, and an anomaly diagnosis and compensation module. The carrier voltage modulation module is used to determine the voltage zero-crossing reference time based on the real-time frequency of the power grid, generate a theoretical delay angle by combining the initial setting command and the remedial control command, obtain the trigger time offset based on the theoretical delay angle, and drive the power switching device to conduct to form a carrier voltage waveform. The instruction parsing module is used to determine the local zero-crossing reference time and waveform distortion point based on the carrier voltage waveform, to calculate the actual resolution angle, to determine the restored control semantics based on the actual resolution angle, and to generate the air conditioning control interface signal based on the restored control semantics. The status monitoring module is used to collect the controlled current sampling sequence and controlled voltage sampling sequence after the air conditioning control interface signal is triggered, extract the harmonic feature set and construct the harmonic fingerprint feature vector, and generate the actual operating status matrix by combining the angle constraint interval corresponding to the control action. The anomaly diagnosis and compensation module is used to calculate the consistency confidence level based on the initial setting command and the actual operating state matrix, obtain the anomaly type flag based on the consistency confidence level and harmonic feature set, and generate a remedial control flow for feedback based on the anomaly type flag.

Images