Railway locomotive air conditioner low harmonic frequency conversion power supply system and control method

Abstract

The application discloses a low-harmonic frequency conversion power supply system and a control method for a rail locomotive air conditioner, relates to the technical field of harmonic control, and determines a harmonic sensitive section by analyzing historical harmonic fault data, determines a plurality of control strategies based on the harmonic sensitive section, starts the locomotive air conditioner, and collects cold start load data of the locomotive air conditioner in real time, pre-processes the cold start load data, and obtains a standard cold start data set, extracts a cold start load index in the standard cold start data set, extracts harmonic trend data, a harmonic trend index and a frequency output function in the standard cold start data set when judging a cold start load abnormality based on the cold start load index, calculates an output comprehensive execution risk value, determines a start risk level based on the output comprehensive execution risk value, selects and executes a corresponding control strategy, and dynamically adjusts a start frequency of the locomotive air conditioner based on the control strategy. The application can greatly improve the start stability of the air conditioner under complex working conditions.

Description

Technical Field

[0001] This invention relates to the field of rail transit technology, specifically to a low-harmonic frequency converter power supply system and control method for rail locomotive air conditioning. Background Technology

[0002] Currently, the air conditioning control systems of rail locomotives commonly employ fixed-frequency start-up strategies or slope-limited linear frequency conversion strategies based on traditional PWM modulation technology. While these methods are structurally simple, during the initial cold start of a train in low-temperature environments, the air conditioning compressor experiences extremely high initial loads, often leading to sudden changes in starting current and severe distortion of the power supply waveform. This results in a large number of high-order harmonics at the output of the frequency converter. These high-order harmonics can easily cause malfunctions in the air conditioning system, frequency output jumps, and even tripping of the protection circuits of the air conditioning system or the entire vehicle's power supply system. Furthermore, existing control systems generally lack the ability to actively identify cold-start load states, cannot dynamically adjust start-up control strategies, exhibit sluggish control response, poor adaptability, and are ill-suited to the complex operating conditions of starting in extremely cold environments. Summary of the Invention

[0003] To address the shortcomings of existing technologies, this invention provides a low-harmonic frequency converter power supply system and control method for railway locomotive air conditioning, solving the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution, comprising the following steps:

[0005] Historical harmonic fault data of locomotive air conditioning is acquired, the historical harmonic fault data is analyzed to determine harmonic sensitive sections, and multiple control strategies are determined based on the harmonic sensitive sections;

[0006] Start the locomotive air conditioner and collect the cold start load data of the locomotive air conditioner in real time during the first time period. Preprocess the cold start load data to obtain a standard cold start dataset.

[0007] Extract the cold start load index from the standard cold start dataset, and perform a comparative evaluation of the cold start load index based on the preset cold start threshold to determine the cold start load status of the locomotive air conditioner.

[0008] When an abnormal cold start load is detected, the harmonic trend data in the standard cold start dataset is extracted, the harmonic trend index is calculated based on the harmonic trend data, and then the frequency output function is calculated based on the harmonic trend index.

[0009] A comprehensive execution risk value is calculated and output based on the cold start load index, harmonic trend index and frequency output function, and the start risk level is determined based on the comprehensive execution risk value.

[0010] Based on the aforementioned risk level, a corresponding control strategy is selected and executed, and the start-up frequency of the locomotive air conditioner is dynamically adjusted during the second time period based on the control strategy.

[0011] Preferably, the harmonic trend data includes total harmonic distortion rate, inverter current fluctuation rate, and inverter voltage fluctuation rate. The harmonic trend index is calculated and output based on the harmonic trend data. The frequency output function is calculated and output based on the obtained harmonic trend index. The comprehensive execution risk value is calculated and output based on the sum of the cold start load index, the harmonic trend index, and the frequency output function.

[0012] Preferably, the risk level of initiation includes a first level, a second level, and a third level, wherein the first level implements a level A control strategy, the second level implements a level B control strategy, and the third level implements a level C control strategy.

[0013] The Class A control strategy is a 3rd-order frequency slow start with an 8-second dwell time for each order. The Class B control strategy is a 5th-order frequency slow start with a 10-second dwell time for each order, and activates an active suppression mechanism. When the harmonic growth rate exceeds 20%, the dwell time of the start frequency is locked, and the total harmonic distortion rate over-limit threshold is set to 18%. The Class C control strategy is a 5th-order frequency slow start with a 15-second dwell time for each order, and immediately freezes the frequency output when the monitored total harmonic distortion rate exceeds 22%.

[0014] Preferably, analyzing the historical harmonic fault data to determine the harmonic-sensitive section includes the following steps:

[0015] The historical harmonic fault data includes the start frequency, the duration of each start frequency, and the interference time point of the harmonic occurrence. A first sequence of each historical harmonic fault data is generated based on the start frequency. The start frequency of the interference time point in the first sequence is located as the first frequency. Various values ​​are statistically analyzed to determine the first occurrence probability of the first frequency in the historical harmonic fault data. The first frequency with the first occurrence probability greater than a first threshold is taken as the second frequency.

[0016] The first sequence in which at least two second frequencies exist is defined as the second sequence, the earliest occurrence time and the latest occurrence time of each second frequency are located, and the allowable delay of each second frequency is calculated based on the earliest occurrence time and the latest occurrence time;

[0017] Multiple frequency combinations are generated based on the second frequency included in each of the second sequences, the frequency combinations are time-aligned based on the allowed delay, and the total number of times each frequency combination is successfully aligned is counted.

[0018] The second occurrence probability of the frequency combination is calculated based on the total number of occurrences. The frequency combination with the second occurrence probability greater than the second threshold is defined as a high-frequency combination. The first frequency with the first occurrence probability greater than the second threshold is taken as the harmonic sensitive segment. The high-frequency combination is verified. The second frequency included in the verified high-frequency combination is taken as the harmonic sensitive segment.

[0019] Preferably, time alignment of the frequency combination includes the following steps:

[0020] The frequency combinations having the same second frequency are used as alignment targets, and the two alignment targets for alignment are defined as a first combination and a second combination. The allowable delay of each second frequency in the first combination and the second combination, as well as the occurrence time period of the second frequency, are obtained. The overlap value range is calculated based on the allowable delay. If the occurrence time period of the second combination can be moved within the overlap value range to be exactly the same as the occurrence time period of the first combination, then the alignment of the first combination and the second combination is determined to be successful.

[0021] Preferably, verifying the candidate frequencies and the high-frequency combination includes the following steps;

[0022] The historical harmonic fault data is divided into an analysis part and a verification part. Based on the historical harmonic fault data in the analysis part, the candidate frequency and the high-frequency combination are obtained. Based on the verification part, the first occurrence count of the candidate frequency and the second occurrence count of the high-frequency combination are obtained. If the first occurrence count of the candidate frequency is greater than a third threshold, the verification is passed. If the second occurrence count of the high-frequency combination is greater than a fourth threshold, the verification is passed.

[0023] Preferably, the control strategy includes multiple start frequencies, and when determining the control strategy, the start frequency of the control strategy is not within the harmonic sensitive section.

[0024] Preferably, the cold start load data includes the vehicle's external ambient temperature, the initial pressure of the refrigerant on the low-pressure side, the condenser surface temperature, the initial current peak change value, the inverter current, and the inverter voltage.

[0025] This invention also provides a low-harmonic frequency converter power supply system for rail locomotive air conditioning, used in the aforementioned control method for the low-harmonic frequency converter power supply of rail locomotive air conditioning. The system includes...

[0026] The acquisition module is used to acquire historical harmonic fault data of the locomotive air conditioner, analyze the historical harmonic fault data to determine the harmonic sensitive section, determine multiple control strategies based on the harmonic sensitive section, and after the locomotive air conditioner is started, the acquisition module acquires the cold start load data of the locomotive air conditioner in real time during the first time period.

[0027] The processing module is used to preprocess the cold start load data to obtain a standard cold start dataset;

[0028] The analysis module extracts cold start load indicators from the standard cold start dataset, performs a comparative evaluation of the cold start load indicators based on a pre-set cold start threshold, and determines the cold start load status of the locomotive air conditioner. When the cold start load is determined to be abnormal, it extracts harmonic trend data from the standard cold start dataset, calculates the harmonic trend index based on the harmonic trend data, calculates the frequency output function based on the harmonic trend index, calculates and outputs a comprehensive execution risk value based on the cold start load indicators, harmonic trend index, and frequency output function, and determines the start risk level based on the comprehensive execution risk value.

[0029] The reconfiguration module selects and executes the corresponding control strategy based on the startup risk level, and dynamically adjusts the startup frequency of the locomotive air conditioner in the second time period based on the control strategy.

[0030] This invention provides a low-harmonic frequency converter power supply system and control method for railway locomotive air conditioning, which has the following beneficial effects:

[0031] This method can accurately quantify the overall load level of the compressor at the start of cold start. After assessing the cold start anomaly, it calculates the harmonic trend index and constructs a dynamic frequency function accordingly. Finally, based on the joint analysis of the cold start load index, harmonic trend index, and frequency output function, it summarizes and generates a comprehensive execution risk value. Using this as the core basis, it performs a secondary risk assessment and dynamic graded response on the start-up strategy of the air conditioning system. This enables the cascading and transition of strategies from rapid standard slow start to multi-stage dwell and real-time harmonic freezing, making the control response more adaptive and safer. It significantly improves the start-up stability of the air conditioning system and the overall safety level of the rail vehicle power supply system under complex conditions such as extreme cold, harmonic pollution, and high load changes. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the low harmonic frequency converter power supply control method for rail locomotive air conditioning according to the present invention;

[0033] Figure 2 This is a schematic diagram of the structure of the low harmonic frequency converter power supply system for rail locomotive air conditioning of the present invention;

[0034] Figure 3 This is a schematic diagram of the evaluation process for the low-harmonic frequency converter power supply control method for rail locomotive air conditioning of the present invention. Detailed Implementation

[0035] 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.

[0036] Please see Figure 1 This invention provides a low-harmonic frequency converter power supply control method for rail locomotive air conditioning. To achieve the above objectives, this invention is implemented through the following technical solution, including the following steps:

[0037] S1. Obtain historical harmonic fault data of locomotive air conditioning, analyze the historical harmonic fault data to determine harmonic sensitive sections, and determine multiple control strategies based on the harmonic sensitive sections.

[0038] S2. Start the locomotive air conditioner and collect the cold start load data of the locomotive air conditioner in the first time period in real time. Preprocess the cold start load data to obtain the standard cold start dataset.

[0039] S3. Extract the cold start load index from the standard cold start dataset, and perform a comparative evaluation of the cold start load index based on the pre-set cold start threshold to determine the cold start load status of the locomotive air conditioner.

[0040] S4. When the cold start load is determined to be abnormal, extract the harmonic trend data from the standard cold start dataset, calculate the harmonic trend index based on the harmonic trend data, and then calculate the frequency output function based on the harmonic trend index.

[0041] S5. Calculate and output the comprehensive execution risk value based on the cold start load index, harmonic trend index and frequency output function, and determine the start-up risk level based on the comprehensive execution risk value.

[0042] S6. Select and execute the corresponding control strategy based on the risk level of startup, and dynamically adjust the startup frequency of the locomotive air conditioner in the second time period based on the control strategy.

[0043] In this embodiment, by installing a sensor array on the rail locomotive and combining it with the Sub-GHz industrial wireless transmission protocol, real-time acquisition and transmission of key load parameters throughout the cold start process are achieved. Subsequently, after receiving the data, the control center uses an STM32 embedded module to perform multi-level preprocessing on the data to obtain a standard cold start dataset. Based on this, the cold start load index is calculated using a load evaluation function and preliminarily compared with a set cold start threshold to determine whether there is any load anomaly during the cold start.

[0044] If the assessment indicates an abnormal operating condition, further features such as real-time THD, current and voltage fluctuation rates are extracted to calculate the harmonic trend index. This index is then combined with cold start load indicators, the harmonic trend index, and the frequency output function to construct a comprehensive execution risk index. Based on this index level, a corresponding three-level slow-start control strategy is triggered to precisely control the start-up path and harmonic suppression strategy, ensuring that the air conditioner can start smoothly and operate with low interference even in extremely cold or complex conditions. Through the above implementation methods, this invention not only achieves joint identification of cold start load and harmonic disturbance states but also establishes a complete process link from perception, assessment, modeling, and control strategy generation, improving environmental adaptability and forward-looking control capabilities for different cold start conditions. By reconstructing the control path driven by the harmonic trend, the peak harmonic value is effectively reduced by more than 40%, improving stability through multi-indicator fusion risk assessment and graded response.

[0045] S2 includes S21 and S22;

[0046] S21. By installing sensor groups at key nodes of the rail locomotive, cold start load data is collected in real time.

[0047] Key components include the outer wall of the railcar, the condenser, the refrigerant circuit, and the air conditioning inverter;

[0048] The sensor group includes a temperature sensor, a refrigerant pressure sensor, a Hall current sensor, a voltage sensor, and a condenser temperature probe. Cold start load data includes the vehicle's ambient temperature, initial refrigerant low-pressure side pressure, condenser surface temperature, initial current peak variation, inverter current, and inverter voltage.

[0049] Cold start load data includes the vehicle's external ambient temperature Tamb, the initial pressure of the refrigerant on the low-pressure side Pref, the condenser surface temperature Tcond, the initial current peak change value △Istart, the inverter current I, and the inverter voltage V;

[0050] S22. Build a control center in the control server of the rail locomotive and set up a short-range industrial-grade wireless sensor network Sub-Ghz. Connect the wireless modules of the sensor group with the air conditioning control center of the rail locomotive and set the data upload frequency to one data packet every 10 seconds to transmit the real-time collected cold start load data to the air conditioning control center.

[0051] In this embodiment, the method deploys sensor groups at key nodes such as the outer wall of the railcar, condenser, refrigerant circuit, and air conditioning inverter to collect cold start load data in real time during the air conditioning cold start phase. A control center is built based on the railcar control server platform, and a short-range industrial-grade wireless sensor network centered on Sub-GHz is established. Various sensor modules with wireless communication capabilities are connected to the air conditioning control center at high speed and with low interference. The data upload frequency is set to transmit one data packet every 10 seconds, achieving stable and real-time transmission of cold start load data. Through the above implementation, not only is the accuracy and update frequency of acquiring pre-cold start operating parameters significantly improved, but a low-power, highly robust, and wiring-free wireless sensing network architecture is also successfully constructed. Its technical objective is to provide a highly consistent, complete, and real-time data foundation for subsequent cold start load level identification, harmonic trend assessment, and frequency control strategy selection, avoiding control errors and instability caused by traditional methods due to parameter acquisition lag, complex wiring, or insufficient sampling resolution.

[0052] In this embodiment, when the control center receives cold start load data in real time, it first preprocesses the cold start load data, including data cleaning, filtering and standardization, to obtain a standard cold start dataset.

[0053] Data cleaning is performed on cold start load data by using upper and lower limit thresholds and change rate removal to remove significant errors, jumps, and physically unreasonable data values.

[0054] The filtering process uses a moving average filter to filter the cold start load data, which removes random jitter and small fluctuations while preserving trend characteristics.

[0055] Standardization is performed by using Z-score standardization to standardize the cold start load data after data cleaning and filtering, thereby eliminating the dimensional influence between all parameters in the cold start load data.

[0056] Step S3 includes S31 and S32;

[0057] S31. Construct a cold start load algorithm model. Extract real-time standard cold start dataset before the cold start of the rail train and input it into the cold start load algorithm model to calculate and output the cold start load index Lload, which quantifies the load index of the air conditioning compressor of the rail train during the cold start stage under the current working conditions.

[0058] The cold start load metric Lload is calculated and output using the following load algorithm model:

[0059] ;

[0060] In the formula, log represents the logarithmic function. n1 represents a small constant to prevent instability caused by division by zero errors or extremely small numbers in the formula; n2 represents the environmental impact coefficient to control the dominance of temperature factors in the model; n3 represents the electrical impact factor weight to control the proportion of initial current jump in the overall load level judgment; n4 represents the thermal state response coefficient to adjust the weight of the effects of the initial pressure Pref on the low-pressure side of the refrigerant and the surface temperature Tcond of the condenser.

[0061] The nonlinear weighting factor representing ambient temperature is used to avoid unreasonable calculation logic caused by negative temperatures in the formula. The absolute value is used to measure the degree of deviation from a warm environment. The exponent is a nonlinear amplification factor.

[0062] This indicates the electrical load; the larger the peak starting current, the greater the rotational resistance.

[0063] This indicates the thermal load. The lower the refrigerant pressure and the colder the condenser, the more severe the cold state and the higher the load.

[0064] The physical meaning of the formula lies in integrating the three factors of cold-state thermal, electrical and environmental into a unified load level evaluation function, directly quantifying the cold start complexity through a mathematical model, and driving the subsequent control logic to adapt dynamically.

[0065] S32. By setting the cold start threshold F1 based on the upper limit of the cold start load of the rail locomotive air conditioning compressor, the load situation of the current rail train cold start is determined. The specific evaluation content is as follows:

[0066] When the cold start load index Lload < cold start threshold F1, it indicates that the cold start load is normal, and the rail locomotive air conditioner can be started normally.

[0067] When the cold start load index Lload is greater than or equal to the cold start threshold F1, it indicates that the cold start load is abnormal, and at this time the harmonic trend-driven frequency ladder reconstruction mechanism is triggered.

[0068] In this embodiment, the method constructs a cold start load algorithm model. Before the cold start of the railcar, it calls a standard cold start dataset matching the current operating conditions from the database and inputs it into the cold start load algorithm model for calculation, outputting the cold start load index Lload. The load index is calculated by an algorithm model that includes parameters such as ambient temperature, initial current peak change, initial refrigerant pressure, and condenser surface temperature. Users are allowed to set a cold start load upper limit as a cold start threshold F1 based on the performance requirements of the onboard air conditioning compressor. The actually calculated load index Lload is then compared with this threshold in real time. Through the above implementation, the railcar air conditioning control has the ability to evaluate complex load conditions in real time during the cold start phase, realizing a closed-loop logic link from "sensing, modeling, and discrimination" to "control triggering." This method can identify the cold start risks caused by high load, high resistance, or extremely cold environments in a timely manner before startup, avoiding current surges, control jumps, or harmonic bursts caused by blind startup, thereby significantly improving the startup safety, intelligent adaptability, and control accuracy of the railcar air conditioning under multiple operating conditions.

[0069] The harmonic trend data in this embodiment includes total harmonic distortion rate, inverter current fluctuation rate, and inverter voltage fluctuation rate. The harmonic trend index is calculated and output based on the harmonic trend data. The frequency output function is calculated and output based on the obtained harmonic trend index. The comprehensive execution risk value is calculated and output based on the cold start load index, harmonic trend index, and frequency output function.

[0070] Specifically, upon determining that the cold start load is abnormal, the standard cold start dataset acquired in real time is extracted, and feature extraction is performed to obtain harmonic trend data. The harmonic trend data includes total harmonic distortion (THD) and inverter current fluctuation rate. and inverter voltage fluctuation rate The total harmonic distortion (THD) is calculated by the ratio of the energy between the higher harmonic components and the fundamental frequency in the inverter voltage V. Here, d represents the differential variable, dt represents the time differential variable, dI represents the differential variable of the inverter current fluctuation rate, and dV represents the differential variable of the inverter voltage fluctuation rate. Then, based on harmonic trend data, the harmonic trend index Htrend is calculated and output to analyze the extreme harmonic fluctuation risk of current rail locomotives during cold starts. The harmonic trend index Htrend is calculated and output using the following algorithm formula:

[0071] ;

[0072] In the formula, THD(t) represents the total harmonic distortion rate at time t, dI(t) represents the differential variable of the inverter current fluctuation rate at time t, dV(t) represents the differential variable of the inverter voltage fluctuation rate at time t, and a1, a2, and a3 represent the inverter current fluctuation rates, respectively. Inverter voltage fluctuation rate The preset weight values ​​for the Total Harmonic Distortion (THD) are set by the user, with a1+a2+a3=1. Finally, based on the obtained harmonic trend index Htrend, the output frequency function Fstep is calculated to dynamically generate the target value of the inverter's output frequency during the current cold start process. The frequency output function Fstep is calculated and output using the following algorithm formula:

[0073] ;

[0074] In the formula, Fstep(t) represents the frequency output function at time t, f0 represents the initial frequency (dimensionless), J represents the total number of frequency steps, and e represents the exponential function. This represents the jump magnitude of the j-th frequency step, set by the controller; a larger value indicates a faster frequency climb. This represents the risk sensitivity at the j-th frequency stage, used to control the influence of the harmonic trend index Htrend. j t represents the start time of the j-th frequency stage. j+1 This indicates the end time of the j-th frequency stage. This represents the time window at time t.

[0075] In this embodiment, the method uses a harmonic trend modeling mechanism to dynamically predict the power quality disturbance risk during cold start, and reconstructs the inverter output frequency control path in real time based on the prediction results, thereby realizing an adaptive soft start control strategy. When the initial comparative evaluation indicates an abnormal cold start load, the harmonic trend-driven frequency ladder reconstruction mechanism is triggered, and the currently collected standard cold start dataset is extracted for feature extraction to obtain harmonic trend data. This harmonic trend data includes three key parameters: first, the total harmonic distortion (THD), which is calculated by the ratio of the higher harmonic components in the inverter voltage signal to the fundamental frequency energy; second, the inverter current fluctuation rate, expressed as the first derivative of the current over time; and third, the inverter voltage fluctuation rate. These parameters comprehensively reflect the power disturbance trend and fluctuation intensity experienced during the cold start process.

[0076] Based on the above three characteristic data, a harmonic trend index Htrend is constructed to quantify the risk level of harmonic disturbances during the current cold start process. This trend index is calculated and output using the following function: Based on the harmonic trend index Htrend calculated at the current moment, a frequency output function Fstep is constructed to achieve dynamic control of the inverter's output frequency target value. Through complete implementation, not only is it capable of accurately modeling and real-time predicting the harmonic disturbance trend during cold start, but it can also proactively reconstruct the frequency ladder based on power quality risks, dynamically adjusting the output frequency rise rate and node settings, thereby effectively suppressing the sudden growth of higher harmonics and reducing current jumps caused by frequency abrupt changes.

[0077] When calculating the comprehensive execution risk value, the comprehensive execution risk value Cexec is calculated and output based on the cold start load index Lload, the harmonic trend index Htrend, and the frequency output function Fstep. The comprehensive execution risk value Cexec is calculated and output using the following algorithm formula to comprehensively analyze the linkage risks of harmonic bursts, load shocks, and control climb failures during the cold start phase of the track.

[0078] ;

[0079] In the formula, This represents the upper limit of the harmonic trend index at time t. dFstep(t) represents the volatility of the frequency output function at time t, i.e. the current frequency rise rate. b1, b2 and b3 represent the preset weight values ​​of the cold start load index Lload, the harmonic trend index Htrend and the frequency output function Fstep, respectively. Their specific values ​​are set by the user, and b1+b2+b3=1. dFstep(t) represents the differential variable of the frequency output function at time t.

[0080] The risk levels are divided into three levels: Level 1, Level 2, and Level 3. Level 1 implements Level A control strategies, Level 2 implements Level B control strategies, and Level 3 implements Level C control strategies.

[0081] The Class A control strategy is a 3rd-order frequency slow start with an 8-second dwell time for each order. The Class B control strategy is a 5th-order frequency slow start with a 10-second dwell time for each order, and activates an active suppression mechanism. When the harmonic growth rate exceeds 20%, the dwell time of the start frequency is locked, and the total harmonic distortion rate over-limit threshold is set to 18%. The Class C control strategy is a 5th-order frequency slow start with a 15-second dwell time for each order, and immediately freezes the frequency output when the total harmonic distortion rate exceeds 22%.

[0082] like Figure 2As shown, when the comprehensive execution risk value Cexec < 4.0, the current cold start risk is classified as Level 1 risk and Level A control strategy is implemented; when 4.0 ≤ comprehensive execution risk value Cexec < 7.5, the current cold start risk is classified as Level 2 risk and Level B control strategy is implemented; when the comprehensive execution risk value Cexec ≥ 7.5, the current cold start risk is classified as Level 3 risk and Level C control strategy is implemented.

[0083] The Class A control strategy initiates a 3-stage slow start, for example, by controlling the cold start frequency path to 15Hz, 30Hz, and 45Hz, progressively advancing the frequency with an 8-second pause at each stage. The Class B control strategy initiates a 5-stage slow start, for example, by controlling the cold start frequency path to 12Hz, 25Hz, 36Hz, 45Hz, and 50Hz, progressively advancing the frequency with a 10-second pause at each stage, and activates an active suppression mechanism. When the harmonic growth rate exceeds 20%, the cold start frequency dwell time is automatically locked, and the total harmonic distortion (THD) over-limit threshold is set to 18%. The Class C control strategy initiates a 7-stage slow start, for example, by controlling the cold start frequency path to 10Hz, 15Hz, 20Hz, 28Hz, 34Hz, 42Hz, and 50Hz, progressively advancing the frequency with a 15-second pause at each stage, and immediately freezes the frequency output when the total harmonic distortion (THD) exceeds 22%.

[0084] In this embodiment, the method uses the cold start load index Lload, harmonic trend index Htrend, and frequency output function Fstep calculated in the previous steps as input variables to conduct a comprehensive summary analysis and construct a comprehensive execution risk value Cexec to reflect the linkage risk level formed by high load, high frequency ramp rate and strong harmonic disturbance during the cold start process.

[0085] This embodiment analyzes historical harmonic fault data to determine harmonic-sensitive sections, including the following steps:

[0086] Historical harmonic fault data includes the starting frequency, the duration of each starting frequency, and the interference time point of the harmonic occurrence. A first sequence of each historical harmonic fault data is generated based on the starting frequency. The starting frequency of the interference time point in the first sequence is located as the first frequency. The first occurrence probability of various values ​​of the first frequency in the historical harmonic fault data is calculated. The first frequency with a first occurrence probability greater than a first threshold is taken as the second frequency.

[0087] Historical harmonic fault data refers to data showing system malfunctions caused by harmonics. This includes the air conditioner's start-up frequency at various time points. A first sequence generated from this historical harmonic fault data is, for example, [15, 17, 19, 21], where the values ​​represent start-up frequencies (units omitted for simplicity). The interference time point corresponds to frequency 19, so 19 is set as the first frequency. If 100 historical harmonic fault data points are collected, and frequency 19 (the first frequency) appears 90 times, the corresponding probability of occurrence is 90 / 100 = 0.9. In this embodiment, the first threshold is set to 0.8. Since 0.9 is greater than 0.8, frequency 19 is set as the second frequency.

[0088] A first sequence with at least two second frequencies is defined as a second sequence. The earliest and latest occurrence times of each second frequency are located, and the allowable delay for each second frequency is calculated based on the earliest and latest occurrence times.

[0089] Based on the above method, a first sequence may contain two or more second frequencies. For example, in the above example, frequencies 17 and 19 are both determined to be second frequencies, so the above first sequence is defined as a second sequence. The earliest and latest occurrence times of the second frequencies in each second sequence are obtained; the occurrence time is the frequency's position in the first sequence. Combining the previously introduced first sequences, for frequency 17, its occurrence time is 2; if there is also a first sequence [13, 15, 17, 19, 21], its occurrence time is 3. Finally, combining all the first sequences, the minimum and maximum occurrence times of the frequencies are determined and used as the earliest and latest occurrence times, respectively. Then, the allowable delay for each second frequency is calculated based on the earliest and latest occurrence times. The following example illustrates the calculation process for the allowable delay, which is calculated using the following formula. .in, To allow for delays, The latest time of occurrence. The earliest occurrence time is used. For example, for frequency A, if its earliest occurrence time is 2 and its latest occurrence time is 6, then its allowable delay is 2. In particular, if the allowable delay is not an integer, it is rounded.

[0090] Multiple frequency combinations are generated based on the second frequency included in each second sequence, the frequency combinations are time-aligned based on the allowed delay, and the total number of times each frequency combination is successfully aligned is counted.

[0091] Specifically, frequency combinations with the same second frequency are used as alignment targets, and two alignment targets for alignment are defined as the first combination and the second combination. The allowable delay of each second frequency in the first combination and the second combination, as well as the occurrence time period of the second frequency, are obtained. The overlap value range is calculated based on the allowable delay. If the occurrence time period of the second combination can be moved within the overlap value range to be exactly the same as the occurrence time period of the first combination, then the alignment of the first combination and the second combination is determined to be successful.

[0092] For example, if second sequence 1 includes second frequencies A and B, and second sequence 2 includes second frequencies A and B, then second sequence 1 and second sequence 2 are used as alignment targets. Second frequencies A and B in second sequence 1 are considered the first combination, and second frequencies A and B in second sequence 2 are considered the second combination. The overlap range is the smaller of the allowed delays of the two second frequencies in the combination. For second frequencies A and B, the calculated allowed delays are 1 and 2 respectively, so the overlap range is 1. For another example, for second frequencies A and B, the calculated allowed delays are 3 and 5 respectively, so the overlap range is 3. The alignment process is described below. For second sequence 1, the occurrence times of second frequencies A and B are 2 and 4 respectively, and for second sequence 2, the occurrence times of second frequencies A and B are 3 and 5 respectively. The overlap range is 1. Therefore, shifting the occurrence times of second frequencies A and B in second sequence 2 to the left by one time unit will make them the same as the occurrence times of second frequencies A and B in second sequence 1. This determines that the first combination and the second combination can be aligned.

[0093] The second occurrence probability of frequency combinations is calculated based on the total number of occurrences. Frequency combinations with a second occurrence probability greater than a second threshold are defined as high-frequency combinations. The first frequency with a first occurrence probability greater than the second threshold is taken as a harmonic sensitive segment. The high-frequency combinations are verified, and the second frequency included in the verified high-frequency combinations is taken as a harmonic sensitive segment.

[0094] This embodiment verifies the candidate frequencies and high-frequency combinations by including the following steps;

[0095] This embodiment verifies high-frequency combinations by including the following steps;

[0096] Historical harmonic fault data is divided into an analysis section and a verification section. High-frequency combinations are obtained based on the historical harmonic fault data in the analysis section, and the occurrence frequency of high-frequency combinations is obtained based on the verification section. If the occurrence frequency of a high-frequency combination is greater than a third threshold, the verification is successful.

[0097] Analyzing 100 first sequences, it is determined that for a certain frequency combination (including second frequency A and second frequency B), 30 second sequences can be aligned, i.e., the total number of occurrences is 30. The probability of the second occurrence is calculated based on the total number of occurrences and the number of first sequences, resulting in a probability of 30 / 100 = 0.3. In this embodiment, the second threshold is set to 0.9. After calculating the probability of the second occurrence for each frequency combination, the previously calculated probability of the first occurrence of the second frequency is compared with the second threshold. If the second threshold is greater than the first threshold (e.g., 0.9), the first frequencies greater than the second threshold are designated as harmonic sensitive areas. For example, if the probability of the first occurrence of frequency 17 is greater than the second threshold, it indicates that harmonic interference is likely to occur within the frequency 17 range. For high-frequency combinations that pass verification, it is shown that when the second frequency combinations within them occur simultaneously, there is a high probability of harmonic interference; therefore, the included second frequencies are designated as harmonic sensitive areas.

[0098] The third threshold is determined based on the amount of historical harmonic fault data included in the analysis and verification sections. This invention divides the historical harmonic fault data into an analysis section and a verification section. After determining the high-frequency combinations in the analysis section, the verification section verifies them, thereby further ensuring the reliability of the analysis results.

[0099] The control strategy includes multiple start frequencies. When determining the control strategy, the start frequency of the control strategy is not within the harmonic sensitive section.

[0100] This invention reduces the risk of harmonic occurrence by pre-determining harmonic-sensitive sections, which are areas where harmonics are prone to occur, and by avoiding these harmonic-sensitive sections when formulating control strategies.

[0101] Please see Figure 3 The present invention also provides a low-harmonic frequency converter power supply system for rail locomotive air conditioning, used in the above-mentioned low-harmonic frequency converter power supply control method for rail locomotive air conditioning, the system comprising:

[0102] The acquisition module is used to acquire historical harmonic fault data of the locomotive air conditioner, analyze the historical harmonic fault data to determine the harmonic sensitive section, determine multiple control strategies based on the harmonic sensitive section, and after the locomotive air conditioner is started, the acquisition module collects the cold start load data of the locomotive air conditioner in real time during the first time period.

[0103] The processing module is used to preprocess the cold start load data to obtain a standard cold start dataset.

[0104] The analysis module extracts cold start load indicators from the standard cold start dataset, performs a comparative evaluation of the cold start load indicators based on a pre-set cold start threshold, and determines the cold start load status of the locomotive air conditioner. When the cold start load is determined to be abnormal, it extracts harmonic trend data from the standard cold start dataset, calculates the harmonic trend index based on the harmonic trend data, and then calculates the frequency output function based on the harmonic trend index. Based on the cold start load indicators, harmonic trend index, and frequency output function, it calculates and outputs a comprehensive execution risk value, and determines the start risk level based on the comprehensive execution risk value.

[0105] The reconfiguration module selects and executes the corresponding control strategy based on the startup risk level, and dynamically adjusts the startup frequency of the locomotive air conditioner in the second time period based on the control strategy.

[0106] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

Claims

1. A method for controlling low-harmonic frequency converter power supplies for rail locomotive air conditioning, characterized in that: Includes the following steps: Acquire historical harmonic fault data of locomotive air conditioning, analyze the historical harmonic fault data to determine harmonic sensitive sections, and determine multiple control strategies based on the harmonic sensitive sections; Start the locomotive air conditioner and collect the cold start load data of the locomotive air conditioner in the first time period in real time. Preprocess the cold start load data to obtain a standard cold start dataset. Extract cold start load indicators from the standard cold start dataset, and compare and evaluate the cold start load indicators based on the pre-set cold start threshold to determine the cold start load status of the locomotive air conditioner. When an abnormal cold start load is detected, harmonic trend data is extracted from the standard cold start dataset. The harmonic trend index is calculated based on the harmonic trend data, and then the frequency output function is calculated based on the harmonic trend index. The comprehensive execution risk value is calculated and output based on the cold start load index, harmonic trend index and frequency output function, and the start-up risk level is determined based on the comprehensive execution risk value. Based on the risk level, select and execute the corresponding control strategy, and dynamically adjust the start frequency of the locomotive air conditioner in the second time period based on the control strategy.

2. The method according to claim 1, characterized in that: Harmonic trend data includes total harmonic distortion rate, inverter current fluctuation rate, and inverter voltage fluctuation rate. Based on the harmonic trend data, the harmonic trend index is calculated and output. Based on the obtained harmonic trend index, the frequency output function is calculated and output. Based on the cold start load index, harmonic trend index, and frequency output function, the comprehensive execution risk value is calculated and output.

3. The method according to claim 1, characterized in that: The risk levels are divided into three levels: Level 1, Level 2, and Level 3. Level 1 implements Level A control strategy, Level 2 implements Level B control strategy, and Level 3 implements Level C control strategy. The Class A control strategy is a 3rd-order frequency slow start with an 8-second dwell time for each order. The Class B control strategy is a 5th-order frequency slow start with a 10-second dwell time for each order, and activates an active suppression mechanism. When the harmonic growth rate exceeds 20%, the dwell time of the start frequency is locked, and the total harmonic distortion rate over-limit threshold is set to 18%. The Class C control strategy is a 5th-order frequency slow start with a 15-second dwell time for each order, and immediately freezes the frequency output when the total harmonic distortion rate exceeds 22%.

4. The method according to claim 1, characterized in that: Analyzing historical harmonic fault data to determine harmonic-sensitive sections includes the following steps: Historical harmonic fault data includes the starting frequency, the duration of each starting frequency, and the interference time point of the harmonic occurrence. A first sequence of each historical harmonic fault data is generated based on the starting frequency. The starting frequency of the interference time point in the first sequence is located as the first frequency. The first occurrence probability of various values ​​of the first frequency in the historical harmonic fault data is calculated. The first frequency with a first occurrence probability greater than a first threshold is taken as the second frequency. A first sequence with at least two second frequencies is defined as a second sequence. The earliest and latest occurrence times of each second frequency are located, and the allowable delay for each second frequency is calculated based on the earliest and latest occurrence times. Multiple frequency combinations are generated based on the second frequency included in each second sequence, the frequency combinations are time-aligned based on the allowed delay, and the total number of times each frequency combination is successfully aligned is counted. The second occurrence probability of frequency combinations is calculated based on the total number of occurrences. Frequency combinations with a second occurrence probability greater than a second threshold are defined as high-frequency combinations. The first frequency with a first occurrence probability greater than the second threshold is taken as a harmonic sensitive segment. The high-frequency combinations are verified, and the second frequency included in the verified high-frequency combinations is taken as a harmonic sensitive segment.

5. The method according to claim 4, characterized in that: Time alignment of frequency combinations includes the following steps: Frequency combinations with the same second frequency are used as alignment targets, and two alignment targets for alignment are defined as the first combination and the second combination. The allowable delay of each second frequency in the first combination and the second combination, as well as the occurrence time period of the second frequency, are obtained. The overlap value range is calculated based on the allowable delay. If the occurrence time period of the second combination can be moved within the overlap value range to be exactly the same as the occurrence time period of the first combination, then the alignment of the first combination and the second combination is determined to be successful.

6. The method according to claim 4, characterized in that, Verification of high-frequency combinations includes the following steps; Historical harmonic fault data is divided into an analysis section and a verification section. High-frequency combinations are obtained based on the historical harmonic fault data in the analysis section, and the occurrence frequency of high-frequency combinations is obtained based on the verification section. If the occurrence frequency of a high-frequency combination is greater than a third threshold, the verification is successful.

7. The method according to claim 1, characterized in that: The control strategy includes multiple start frequencies. When determining the control strategy, the start frequency of the control strategy is not within the harmonic sensitive section.

8. The method according to claim 1, characterized in that, Cold start load data includes ambient temperature outside the vehicle, initial pressure of the refrigerant on the low-pressure side, condenser surface temperature, initial current peak variation, inverter current, and inverter voltage.

9. A low-harmonic frequency converter power supply system for rail locomotive air conditioning, used to implement the low-harmonic frequency converter power supply control method for rail locomotive air conditioning according to any one of claims 1-8, characterized in that: The system includes, The acquisition module is used to acquire historical harmonic fault data of the locomotive air conditioner, analyze the historical harmonic fault data to determine the harmonic sensitive section, determine multiple control strategies based on the harmonic sensitive section, and after the locomotive air conditioner is started, the acquisition module collects the cold start load data of the locomotive air conditioner in real time during the first time period. The processing module is used to preprocess the cold start load data to obtain a standard cold start dataset; The analysis module extracts cold start load indicators from the standard cold start dataset, performs a comparative evaluation of the cold start load indicators based on a pre-set cold start threshold, and determines the cold start load status of the locomotive air conditioner. When the cold start load is determined to be abnormal, it extracts harmonic trend data from the standard cold start dataset, calculates the harmonic trend index based on the harmonic trend data, calculates the frequency output function based on the harmonic trend index, calculates and outputs a comprehensive execution risk value based on the cold start load indicators, harmonic trend index, and frequency output function, and determines the start risk level based on the comprehensive execution risk value. The reconfiguration module selects and executes the corresponding control strategy based on the startup risk level, and dynamically adjusts the startup frequency of the locomotive air conditioner in the second time period based on the control strategy.

Images