Charging and discharging control method and device for elevator energy storage system, medium and product

Abstract

The invention discloses a charging and discharging control method and device for an elevator energy storage system, a medium and a product, and relates to the field of data processing. The method comprises the steps that real-time operation data such as the position and the load capacity of an elevator car are obtained, and the instantaneous power generation power and the power change rate of a traction machine are determined accordingly; determining a power change predicted value in a future preset time period based on the parameters, and obtaining the total amount of power generation energy required to be stored by the energy storage system through zero setting of a positive value, integral operation and efficiency compensation; according to the power change rate, the state of the energy storage unit and the power grid electricity price time period, the charging and discharging priorities fed back by the super capacitor, the lithium battery and the power grid are determined; controlling the bidirectional DC / DC circuit to switch charging and discharging modes according to the priority, and adjusting the PWM duty ratio of the brake resistor; and based on the charging and discharging mode and the duty ratio, charging and discharging control over the elevator energy storage system is achieved. According to the method, the problem of inaccurate real-time matching of the dynamic load and the energy fluctuation of the current elevator energy storage system can be relieved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a charge and discharge control method, device, medium and product for an elevator energy storage system. Background Art

[0002] Traditional elevator technology relies heavily on the grid for power supply, which presents significant drawbacks. For example, elevators experience significant load fluctuations and frequent starts and stops, which can lead to sudden increases in grid load during peak hours, causing voltage fluctuations and other issues, impacting power supply stability and the operation of surrounding equipment. Furthermore, regenerative energy generated during braking or light-load upward travel cannot be effectively recovered and is mostly converted into wasted heat through braking resistors. This increases cooling costs, leads to low energy utilization, and drives up electricity costs over the long term.

[0003] To address these issues, energy storage systems are being introduced to optimize elevator energy management. These systems, through power regeneration devices or components like supercapacitors and batteries, store the regenerative energy generated during elevator operation, releasing it prioritizing energy consumption during equipment consumption, thereby reducing reliance on the power grid. This energy recycling approach can significantly reduce energy consumption, smooth grid load fluctuations, lower electricity costs, and reduce cooling requirements for the equipment room.

[0004] However, current elevator energy storage systems still face the bottleneck of inaccurate real-time matching of dynamic loads with energy fluctuations. Elevator loads and energy demands change dynamically, and existing control strategies struggle to respond promptly to instantaneous power changes. This can lead to improper charging and discharging timing, resulting in energy waste or insufficient power supply. Summary of the Invention

[0005] In response to the above-mentioned technical problems and defects, the purpose of the present invention is to provide a charging and discharging control method, equipment, medium and product for an elevator energy storage system, which can alleviate the problem of inaccurate real-time matching between dynamic load and energy fluctuations in current elevator energy storage systems.

[0006] To achieve the above-mentioned purpose, in a first aspect, the present invention provides a charge and discharge control method for an elevator energy storage system, comprising: obtaining real-time operation data of the elevator, wherein the real-time operation data includes the elevator car position, load capacity, traction machine speed, DC bus voltage and current; determining the instantaneous generated power and power change rate of the elevator traction machine according to the real-time operation data; determining a power change prediction value of the elevator traction machine in a future preset time period based on the instantaneous generated power and power change rate; determining the total amount of generated energy that the elevator energy storage system needs to store in the future preset time period based on the power change prediction value; determining the total amount of generated energy that the elevator energy storage system needs to store in the future preset time period based on the power change rate, the energy storage unit status of the elevator energy storage system, and the power grid voltage ...; determining the total amount of generated energy that the elevator energy storage system needs to store in the future preset time period; determining the total amount of generated energy that the elevator energy storage system needs to store in the future preset time period; determining the total amount of generated energy that the elevator energy storage system needs to store in the future preset time period; determining the total amount of generated energy that the elevator energy storage system needs to store in the future preset time period; determining the total amount of During a charging time period, the charging and discharging priority of the supercapacitor, lithium battery or grid feedback of the elevator energy storage system is determined, and the energy storage unit status includes the charge state of the supercapacitor and the health state of the lithium battery; according to the charging and discharging priority, the bidirectional DC / DC circuit of the elevator energy storage system is controlled to switch the charging and discharging mode, and the PWM duty cycle of the braking resistor in the elevator energy storage system is adjusted. The braking resistor is used to consume the regenerative electric energy generated during the elevator electrical braking process, and the PWM duty cycle is used to control the power-on time ratio of the braking resistor to adjust the power consumption of the braking resistor; based on the charging and discharging mode and the PWM duty cycle of the braking resistor, the elevator energy storage system is controlled to charge and discharge.

[0007] Optionally, in some embodiments, determining the power change prediction value of the elevator traction machine within a future preset time period based on the instantaneous generated power and the power change rate includes: generating a multi-dimensional feature data set based on the real-time operation data and the historical power change rate database; inputting the multi-dimensional feature data set into a pre-trained XGBoost regression model to obtain an initial power change prediction curve containing power prediction values ​​at different time points within a future preset time period; and performing error compensation on the initial power prediction curve through a sliding time window based on the real-time detected traction machine vibration spectrum data to obtain the power change prediction value of the elevator traction machine.

[0008] Using the technical solutions of the above-mentioned embodiments, an XGBoost regression model combined with vibration spectrum error compensation enables high-precision prediction of elevator traction motor power changes. A multi-dimensional feature dataset encompasses operating status and historical data, and a sliding time window compensates for vibration interference (such as power measurement errors caused by car sway). This provides a reliable basis for pre-planning charging and discharging strategies for the energy storage system, improves energy recovery efficiency, and adapts to dynamic power changes under complex operating conditions such as elevator start-stop, acceleration, and deceleration, ensuring timely and accurate energy storage control.

[0009] Optionally, in some embodiments, determining the total amount of generated energy that needs to be stored in the elevator energy storage system within the future preset time period based on the power change prediction value includes: setting the positive values ​​in the power change prediction value to zero and retaining the negative values ​​to obtain a generated power sequence; performing an integration operation on the generated power sequence within the future preset time period to obtain a theoretically storable amount of electricity; and performing loss compensation on the theoretically storable amount of electricity based on the real-time charging and discharging efficiency of the energy storage unit in the elevator energy storage system to obtain the actual total amount of generated energy that needs to be stored.

[0010] The technical solutions of the above embodiments are used to denoise the generated power sequence (retaining negative power, i.e., the regenerative power portion) and accurately calculate the actual storable power through charge and discharge efficiency compensation (taking into account the charge and discharge losses of supercapacitors and lithium batteries). This avoids deviations between theoretical and actual values, optimizes energy distribution in the energy storage system, reduces ineffective charging and discharging (such as energy wasted during grid feedback), improves energy storage unit utilization, and reduces system energy consumption, meeting the requirements of elevator energy-saving design.

[0011] Optionally, in some embodiments, the charging and discharging priority of the supercapacitor, lithium battery or grid feedback of the elevator energy storage system is determined based on the power change rate, the state of the energy storage unit of the elevator energy storage system and the grid electricity price period, including: determining the fluctuation state of the current load of the elevator traction machine based on the power change rate, the fluctuation state includes high-frequency transient fluctuations or steady-state fluctuations; determining the availability of the energy storage unit according to the charge state of the supercapacitor and the health state of the lithium battery; determining the current electricity price period based on the real-time clock and the electricity price database, the electricity price period includes a peak period, a valley period or a normal period; inputting the fluctuation state, the availability of the energy storage unit and the current electricity price period into a preset rule base to obtain the charging and discharging priority.

[0012] Adopting the technical solutions of the above embodiments, dynamic charging and discharging priorities are generated based on multi-dimensional decision-making based on load fluctuations, energy storage status, and electricity price periods. This achieves coordinated control with "fast transient response (millisecond-level charging and discharging of supercapacitors), high steady-state energy efficiency (long-term energy storage of lithium batteries), and low electricity costs." This extends the life of lithium batteries, improves supercapacitor utilization, and adapts to the energy management needs of all elevator operating conditions (light load / heavy load, high speed / low speed).

[0013] Optionally, in some embodiments, controlling the bidirectional DC / DC circuit of the elevator energy storage system to switch the charge and discharge mode according to the charge and discharge priority, and adjusting the brake resistor PWM duty cycle of the elevator energy storage system includes: generating a switching signal of the charge and discharge mode according to the charge and discharge priority, the charge and discharge mode including a supercapacitor priority charging mode, a lithium battery priority discharging mode or a grid feedback mode; based on the switching signal, controlling the bidirectional DC / DC circuit to switch the charge and discharge mode; after switching the charge and discharge mode, determining the brake resistor PWM duty cycle according to the detected real-time deviation of the bus voltage.

[0014] Adopting the technical solutions of the above-mentioned embodiments, the energy storage system adapts to its operating conditions by prioritizing the bidirectional DC / DC circuit to switch between charge and discharge modes and adjusting the brake resistor duty cycle based on the bus voltage. This enables millisecond-level response for transient power generation using supercapacitors, long-term storage using lithium batteries, and dynamic adjustment of the brake resistor duty cycle. This solution integrates the characteristics of multiple energy storage units, optimizes energy distribution, reduces power device losses, improves the energy efficiency and reliability of elevator energy storage systems, and adapts to real-time energy management in complex scenarios such as starting, stopping, and accelerating / decelerating.

[0015] Optionally, in some embodiments, determining the PWM duty cycle of the braking resistor based on the detected real-time deviation of the bus voltage includes: constructing a sliding mode surface in a sliding mode control algorithm based on the real-time deviation and change rate of the bus voltage; when the absolute value of the sliding mode surface is greater than a sliding mode threshold, performing a sliding mode control calculation to obtain a first duty cycle increment; performing a gradient constraint on the first duty cycle increment based on the real-time temperature of the braking resistor to obtain a second duty cycle increment; and generating the PWM duty cycle of the braking resistor based on the second duty cycle increment, the real-time temperature of the braking resistor, and the historical action frequency.

[0016] Adopting the technical solution of the above embodiment, a sliding mode control algorithm is used to construct a sliding mode surface based on the bus voltage deviation and rate of change. Threshold triggering and temperature constraints are combined to achieve robust regulation of the brake resistor duty cycle. The rapid convergence of sliding mode control ensures that the bus voltage stabilizes quickly during elevator braking. Fractional-order differentials capture non-stationary voltage fluctuations, and the time-varying coupling coefficient dynamically optimizes the control gain. Temperature gradient constraints protect the brake resistor from overheating and failure. This solution maintains small voltage deviation fluctuations even in strong disturbance scenarios, providing a high-precision, low-jitter solution for bus voltage control in elevator electrical braking, improving system stability and power device life.

[0017] Optionally, in some embodiments, the functional expression of the sliding surface includes: Among them, S represents the value of the sliding surface, ΔV represents the real-time deviation of the bus voltage, represents the fractional differential term of ΔV, α is the order of the fractional differential operator, t is time, λ(t) represents the time-varying coupling coefficient, λ0 is the initial coupling coefficient, sign(ΔV) represents the sign function based on ΔV, γ is the potential field intensity coefficient, β is the potential field attenuation coefficient, τ represents the integral time variable, k is the cumulative gain coefficient, and e is a natural constant.

[0018] Using the technical solutions of the above embodiments, the sliding mode surface function overcomes the limitations of traditional sliding modes by integrating fractional-order differentials, a time-varying coupling coefficient λ(t), and an exponentially decaying potential field term. The fractional-order operator captures the slowly varying voltage trend, the time-varying mechanism adapts to sudden load changes, and the potential field correction reduces high-frequency switching. This provides a high-precision, low-loss energy regulation solution for elevator energy storage systems, adaptable to complex operating scenarios.

[0019] In a second aspect, an embodiment of the present invention provides an electronic device, comprising: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the electronic device to execute the method described in the first aspect or the second aspect, and any possible implementation of the first aspect or the second aspect.

[0020] In a third aspect, the present invention provides a computer-readable storage medium comprising instructions, which, when executed on the electronic device, enables the electronic device to execute the method described in the first aspect or the second aspect, and any possible implementation of the first aspect or the second aspect.

[0021] In a fourth aspect, the present invention provides a computer program product comprising instructions, which, when the computer program product is run on the electronic device, enables the electronic device to execute the method described in the first aspect or the second aspect, and any possible implementation of the first aspect or the second aspect.

[0022] It is understood that the electronic device provided in the second aspect, the storage medium provided in the third aspect, and the computer program product provided in the fourth aspect are all used to execute the method provided by the present invention. Therefore, the beneficial effects that can be achieved can be referred to the beneficial effects of the corresponding methods and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 1 is a schematic diagram of the architecture of an elevator energy storage system according to an embodiment of the present invention; Figure 2 1 is a flow chart of a charge and discharge control method for an elevator energy storage system according to an embodiment of the present invention; Figure 3 It is a schematic diagram of the architecture of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] The terms used in the following embodiments of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. As used in the specification of the present invention, the singular expressions "a," "an," "above," "the," and "this" are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" as used in the present invention refers to any and all possible combinations of one or more of the listed items.

[0025] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying relative importance or implicitly indicating the quantity of the technical features indicated. Thus, a feature designated "first" or "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of the present invention, unless otherwise specified, "plurality" means two or more.

[0026] It should also be noted that, unless otherwise clearly specified and limited, in the embodiments of the present invention, terms such as "setting" and "connection" should be understood in a broad sense. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, or it can be the internal connection of two components; it can be a wired communication connection or a wireless communication connection. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to the specific circumstances. The embodiments of the present invention are described in detail below.

[0027] In the related art, conventional methods for controlling the charging and discharging of elevator energy storage systems are usually based on voltage thresholds or power balance logic. Specifically, when the elevator brakes and generates regenerative energy, if the DC bus voltage exceeds the preset upper limit, the control system triggers the braking resistor to start working, consuming energy through the resistor to suppress the voltage increase; if the voltage is lower than the preset lower limit, the energy storage device (such as a battery or capacitor) is started to release energy to maintain voltage stability. During the charging and discharging process, the input and output of the braking resistor are mostly controlled by switch quantities, and the adjustment strategy is relatively fixed, and does not fully combine dynamic parameters such as the power change rate and the state of the energy storage unit (such as remaining capacity, SOC). In addition, the control of the DC / DC circuit often lags behind the action of the braking resistor, and lacks a pre-response mechanism. The system is prone to voltage overshoot or adjustment delay when the power suddenly changes, and the control parameters are usually fixed preset values, which are difficult to adapt to the energy fluctuation characteristics under different working conditions.

[0028] Therefore, an embodiment of the present invention provides a charge and discharge control method for an elevator energy storage system. By obtaining real-time operating data such as the elevator car position and load capacity, the instantaneous generated power and power change rate of the traction machine are calculated. Based on this, the power change within a preset time period in the future is predicted and the total amount of generated energy that the energy storage system needs to store is determined. The charge and discharge priority of the supercapacitor, lithium battery, or grid feedback is determined in combination with the power change rate, supercapacitor charge state, lithium battery health status, and grid electricity price period. According to the priority, the bidirectional DC / DC circuit is controlled to switch the charge and discharge mode and the brake resistor PWM duty cycle is adjusted to control its power-on time ratio, ultimately realizing the charge and discharge control of the energy storage system.

[0029] The dynamic power prediction mechanism in this process can predict energy fluctuations in advance, solving the regulation delays or voltage overshoots caused by voltage threshold control in traditional methods. The priority decision-making based on multi-parameter fusion takes into account the energy storage unit status and electricity price period, changing the limitations of traditional single voltage control. For example, when the supercapacitor is highly charged, other energy storage or feedback paths are prioritized to avoid unnecessary use of the braking resistor. The dynamic adjustment of the braking resistor PWM duty cycle enables refined control of regenerative energy consumption. Compared with traditional fixed duty cycle or on-off control, it can smooth energy fluctuations by adjusting the power-on time ratio when the power rate changes significantly, reducing the risk of voltage mutation. The coordinated control of the bidirectional DC / DC circuit and the braking resistor synchronizes the hardware response with the energy allocation strategy, solving the problem of DC / DC circuit action lagging behind energy demand in traditional methods. In-depth utilization of the supercapacitor charge state and lithium battery health status can prevent overcharging and over-discharging of energy storage units, extending equipment life. The grid electricity price period is included in the charging and discharging priority decision-making, enabling energy storage charging during off-peak periods and energy storage discharging during peak periods, reducing the overall electricity cost of the elevator system.

[0030] This embodiment forms a closed-loop control of the entire process of energy prediction, allocation, and consumption through the above-mentioned mechanism, and comprehensively optimizes the dynamic response capability, energy utilization efficiency, equipment protection, and economy, effectively solving the problems of inaccurate real-time matching of dynamic load and energy fluctuation, lagging parameter adjustment, and failure to integrate multiple influencing factors in related technologies.

[0031] like Figure 1 As shown, in this embodiment, the elevator energy storage system mainly consists of the following parts: Energy storage unit: includes supercapacitors and lithium batteries. Supercapacitors have high power density characteristics and are used to quickly absorb or release high-frequency regenerative energy generated by elevator braking. Their state of charge (SOC) serves as a key parameter for charging and discharging priority decisions. Lithium batteries have high energy density and are used to store medium- and long-term energy. Their state of health (SOH) is incorporated into the system control strategy to avoid overcharging and over-discharging.

[0032] Bidirectional DC / DC circuit: Connects the energy storage unit to the DC bus and enables bidirectional energy flow between the energy storage unit and the power grid or elevator system by switching the charge and discharge modes (such as buck charging and boost discharging). Its operating mode is controlled by the charge and discharge priority, and the voltage and current can be dynamically adjusted according to energy demand.

[0033] Braking resistor: Connected in parallel with the DC bus, it adjusts the proportion of power-on time through the PWM duty cycle. It is used to consume the excess regenerative energy that cannot be absorbed by the energy storage unit during the elevator's electrical braking process, plays a role in voltage stability and energy buffering, and prevents the DC bus voltage from being too high.

[0034] Sensor module: used to collect real-time elevator operation data, including car position, load, traction motor speed, DC bus voltage and current, and monitor the status of energy storage units (such as supercapacitor SOC and lithium battery SOH) to provide real-time parameter input for system control.

[0035] Control processing module: Based on the data collected by sensors, it calculates the instantaneous power generation and power change rate of the traction machine, predicts future power change trends, and determines the charging and discharging priority based on the energy storage unit status and grid electricity price period. It then controls the mode switching of the bidirectional DC / DC circuit and the PWM duty cycle of the braking resistor, achieving full-process closed-loop control of the energy storage system.

[0036] The various components of the elevator energy storage system in this embodiment work together through electrical connections and data interaction, forming a complete link of "energy prediction-priority decision-hardware execution", which not only meets the needs of dynamic energy management during elevator operation, but also optimizes energy recovery efficiency, system stability and economy.

[0037] The following combination Figure 2 , taking the control processing module as the execution body, a charge and discharge control method for an elevator energy storage system of this embodiment is described in detail, specifically comprising the following steps: Step 201: Acquire real-time operation data of the elevator.

[0038] The real-time operation data includes the elevator car position, load capacity, traction machine speed, DC bus voltage and current.

[0039] The elevator car position refers to the real-time height coordinate of the car in the shaft; the load capacity is the total mass of passengers and cargo in the car; the traction motor speed is the angular velocity of the motor shaft that drives the elevator; the DC bus voltage is the voltage value of the DC power supply network between the energy storage system and the motor driver; the DC bus current reflects the energy transmission intensity between the energy storage unit and the drive system.

[0040] Specifically, the position of the car is obtained through a rotary encoder installed at the end of the traction machine shaft. The encoder generates a pulse signal as the traction machine rotates, and the system converts it into the actual position of the car through pulse counting and pulse equivalent; the load weight is detected by a weighing sensor installed at the bottom of the car or the load-bearing structure of the traction machine. The sensor converts the weight into an analog electrical signal through strain gauges or piezoresistive effect, which is input into the control system after amplification and filtering; the traction machine speed is calculated by the pulse frequency of the rotary encoder and converted into a speed value by the number of pulses per unit time; the DC bus voltage is collected by a Hall voltage sensor or a voltage divider resistor network. The Hall sensor measures the voltage non-contactly through the principle of electromagnetic induction, and the voltage divider resistor detects the voltage signal after resistance division; the DC bus current is collected by a Hall current sensor or a shunt. The Hall sensor detects the magnetic field generated by the current based on Ampere's law, and the shunt calculates the current by detecting the voltage drop at both ends and combining the resistance value. The analog signals output by the above sensors are converted into digital signals through the analog-to-digital conversion module (ADC), or directly input into the control processing module through a digital interface (such as the pulse signal of an incremental encoder). At the same time, the control processing module realizes data synchronization through a real-time bus (such as CAN, RS485) or hard-wired connection to ensure that the acquisition frequency meets the real-time monitoring requirements of the elevator operation status, providing an accurate data basis for subsequent power calculation and charge and discharge control.

[0041] Step 202: Determine the instantaneous generated power and power change rate of the elevator traction machine based on the real-time operating data.

[0042] Specifically, the control processing module determines the elevator motor's instantaneous generated power and power change rate based on real-time operating data. For instantaneous generated power, the module first calculates the motor's output mechanical power based on the motor's speed and torque characteristic curves, combined with real-time collected motor speed data. Simultaneously, the DC bus-side electrical power is calculated by multiplying the DC bus voltage and current in real time.

[0043] In the power generation state, taking into account the energy conversion efficiency, the mechanical power is multiplied by the motor power generation efficiency coefficient (usually calibrated through experiments) and cross-validated with the DC bus power. The weighted average of the two is taken as the final instantaneous power generation power, and the weight coefficient is dynamically adjusted according to the operating conditions.

[0044] For the power change rate, the module uses a sliding window difference algorithm to perform linear regression analysis on multiple continuously collected instantaneous power generation samples (such as data from the last five sampling cycles), calculates the ratio of the power difference between adjacent samples to the time interval, and then uses the Kalman filter algorithm to reduce the noise of the calculation results to eliminate the influence of sensor measurement noise, thereby obtaining a smooth and accurate power change rate, providing a reliable basis for subsequent predictions.

[0045] Step 203: Based on the instantaneous generated power and the power change rate, a predicted value of the power change of the elevator traction machine within a preset time period in the future is determined.

[0046] Specifically, the control processing module determines the power change forecast value within a preset time period in the future through a hybrid prediction model based on the determined instantaneous generated power and power change rate. First, the autoregressive integrated moving average model (ARIMA) is used to perform a time series analysis on the historical power data to capture the periodicity and trend characteristics of power changes. At the same time, combined with the physical model of elevator operation, real-time operating data such as car position and load capacity are input into the model to calculate the potential energy and kinetic energy changes of the elevator within the preset time period in the future, and then deduce the corresponding power demand. Then, a weighted fusion algorithm is used to fuse the prediction results of the two models, and the weights are dynamically adjusted according to the prediction accuracy of the model under the current working conditions.

[0047] To cope with sudden operating conditions, a Markov chain can also be introduced to predict the power state transition probability. When the power change rate is detected to exceed the preset threshold, an emergency prediction correction mechanism is triggered. The prediction accuracy is improved by increasing the sampling frequency and shortening the prediction step size, ensuring that the prediction results can promptly reflect the sudden changes in the elevator operating status and provide accurate guidance for the energy distribution of the energy storage system.

[0048] In some embodiments, the specific implementation method of using Markov chain to predict the power state transition probability is as follows: first, the state space of the generated power of the elevator traction machine is divided, and the power variation range is determined based on the statistical analysis of historical operation data, and it is divided equally into N state intervals (such as [-∞, -P1) is state S1, [-P1, -P2) is state S2,…, [Pn, +∞) is state SN).

[0049] Then, a state transition matrix is ​​constructed. The number of transitions between states is counted through a sliding time window (such as the last 100 sampling cycles). The frequency of transitions from state Si to state Sj is calculated. After Laplace smoothing, the transition probability Pij is obtained to form an N×N transition matrix P.

[0050] To improve the prediction accuracy, a time-varying Markov chain model is adopted to dynamically adjust the transfer matrix according to the current elevator operating conditions (such as load and speed). The operating condition identification is realized through the K-nearest neighbor algorithm, which matches the current operating parameters with the historical operating condition template library and selects the transfer matrix corresponding to the most similar operating condition.

[0051] During the prediction phase, the probability distribution of transitions to various states within the next M time steps is calculated based on the current power state Si and the transition matrix P. The maximum a posteriori probability criterion is used to determine the most likely power state sequence. To cope with unexpected operating conditions, when the power change rate exceeds a preset threshold, a matrix update mechanism is triggered. An online learning algorithm (such as stochastic gradient descent) is used to quickly adjust the transition probability, allowing the model to adapt to new operating conditions in a timely manner. At the same time, a confidence assessment mechanism is introduced to determine the reliability of the prediction by calculating the entropy value of the predicted state. When the entropy value exceeds the threshold, the Markov chain prediction weight is reduced, and the influence factors of other prediction models (such as the physical model) are increased to ensure the stability and robustness of the prediction results.

[0052] Step 204: Determine the total amount of generated energy that the elevator energy storage system needs to store within the future preset time period based on the power change prediction value.

[0053] Specifically, a time integration process is performed on the power change prediction value within a future preset time period to determine the total amount of generated energy that needs to be stored in the elevator energy storage system.

[0054] First, the prediction time period is divided into multiple equally spaced small time units (such as 10ms), and the instantaneous power prediction value within each time unit is judged: if the power value is positive (power generation state), the power value is multiplied by the length of the time unit to obtain the regenerated energy within the period; if the power value is negative (energy consumption state), the data of this period is ignored (because the energy storage system needs to release energy instead of storing it at this time).

[0055] By accumulating the energy elements under all power generation states, the module obtains a preliminary total regenerative energy. Based on this, the module further adjusts the energy storage system's energy conversion efficiency, for example, taking into account the internal resistance loss of supercapacitors and lithium batteries during charging (typically setting the charging efficiency coefficient to 0.92-0.98). The preliminary total energy is multiplied by the corresponding efficiency coefficient to obtain the actual total amount of storable generated energy.

[0056] Furthermore, the control processing module retrieves the rated capacity data of the energy storage units (such as the maximum storage capacity of supercapacitors and the available capacity of lithium batteries) in real time. If the calculated energy required for storage exceeds the sum of the remaining capacities of the current energy storage units, the remaining capacity is used as the upper limit for truncation to ensure that the calculated energy value does not exceed the actual storage capacity of the system. Throughout this process, dynamic time integration, efficiency correction, and capacity boundary verification are used to accurately calculate the total amount of generated energy required to be stored by the energy storage system, providing a quantitative basis for subsequent charging and discharging strategies.

[0057] Step 205 : Determine the charging and discharging priority of the supercapacitor, lithium battery, or grid feedback of the elevator energy storage system based on the power change rate, the energy storage unit status of the elevator energy storage system, and the grid electricity price period. The energy storage unit status includes the charge state of the supercapacitor and the health state of the lithium battery.

[0058] Specifically, a three-level priority decision-making system is first established: the first level is supercapacitors, which have a millisecond-level charging and discharging response speed and are suitable for high-frequency power fluctuation scenarios; the second level is lithium batteries, which have high energy density but slow response speed and are suitable for medium and low frequency, large-capacity energy storage; the third level is grid feedback or grid power withdrawal, which is significantly affected by the electricity price period.

[0059] In the specific decision-making process, the control processing module first analyzes the power change rate: when the power change rate is greater than the preset high-frequency threshold (such as 5kW / ms), the priority of the supercapacitor is forced to increase to ensure that the instantaneous high-power regenerative energy is quickly absorbed; when the power change rate is in the medium and low frequency range, the coordinated decision-making of the lithium battery and supercapacitor is initiated.

[0060] The control processing module collects the supercapacitor's state of charge (SOC) and lithium battery's state of health (SOH) in real time: If the supercapacitor's SOC is lower than 20% and the lithium battery's SOH is higher than 70%, the supercapacitor is given priority in charging scenarios to leverage its fast response advantage. If the supercapacitor's SOC is higher than 85%, its charging priority is lowered and switched to the lithium battery or grid feedback path. If the lithium battery's SOH is lower than 50%, its deep charging and discharging is restricted, and the supercapacitor or braking resistor is prioritized for energy regulation.

[0061] The control processing module presets a timing strategy for the power grid price period: during off-peak hours (such as 23:00-7:00), even if the energy storage unit is not full, the priority of grid power supply will be reduced, and renewable energy will be used for charging first; during peak hours (such as 9:00-18:00), the energy storage unit energy will be released first to supply power, reducing grid power supply.

[0062] The control processing module constructs a decision table that includes the power change rate interval, SOC / SOH threshold range, and electricity price period rules, and uses a fuzzy logic algorithm to perform weighted calculations on each factor. It ultimately outputs the optimal charging and discharging priority order under the current operating conditions (such as "supercapacitor charging > lithium battery charging > grid feedback" or "lithium battery discharging > supercapacitor discharging > grid power extraction") to ensure that the energy storage system achieves a dynamic balance between safety, efficiency, and economy.

[0063] Step 206 : Control the bidirectional DC / DC circuit of the elevator energy storage system to switch the charge and discharge mode according to the charge and discharge priority, and adjust the PWM duty cycle of the braking resistor in the elevator energy storage system.

[0064] The braking resistor is used to consume the regenerative electric energy generated during the electrical braking process of the elevator, and the PWM duty cycle is used to control the proportion of the power-on time of the braking resistor to adjust the power consumption of the braking resistor.

[0065] Specifically, first analyze the results of the above charge and discharge priority and convert them into specific hardware control instructions. For bidirectional DC / DC circuits, the control processing module determines the operating mode based on the priority: If the priority is "supercapacitor charging", the bidirectional DC / DC circuit is controlled to enter buck mode, converting the DC bus voltage to a charging voltage suitable for the supercapacitor, and accurately adjusting the charging current through a closed-loop PI control algorithm so that the current value follows the preset charging curve (such as the constant current-constant voltage charging curve).

[0066] If the priority is "supercapacitor discharge", it switches to boost mode to increase the supercapacitor voltage to the bus voltage level, and dynamically adjusts the boost ratio to match the load demand.

[0067] The charge and discharge control logic for lithium batteries is similar to that of supercapacitors, but the upper limit of the charge and discharge current will be dynamically adjusted according to the health state (SOH) of the lithium battery. For example, when the SOH is lower than 60%, the maximum charge current will be limited to 50% of the rated current to protect the battery life.

[0068] Regarding the brake resistor PWM duty cycle adjustment, the control processing module dynamically calculates the duty cycle based on the predicted power change and the remaining capacity of the energy storage unit. When the predicted regenerative energy exceeds the energy storage unit's absorptive capacity, the initial duty cycle is calculated using the formula D = k (Pregen - Pstorage) / R, where Pregen is the predicted regenerative power, Pstorage is the energy storage unit's absorptive power, R is the brake resistor value, and k is a safety factor (typically 0.9).

[0069] Subsequently, an adaptive PID control algorithm is used to correct the initial duty cycle in real time, and the bus voltage is used as the feedback quantity. When it is detected that the bus voltage rises too fast, the duty cycle is automatically increased; otherwise, the duty cycle is reduced, forming a "prediction-execution-feedback" closed-loop control.

[0070] In some embodiments, the control processing module will also make secondary adjustments to the duty cycle according to the grid electricity price period: during peak electricity price periods, the duty cycle is appropriately reduced to give priority to using the energy storage unit to store energy; during off-peak electricity price periods, a higher duty cycle is allowed to reduce the demand for energy storage charging.

[0071] Step 207: Control the elevator energy storage system to charge and discharge based on the charge and discharge mode and the brake resistor PWM duty cycle.

[0072] Specifically, the charge / discharge pattern and the brake resistor PWM duty cycle are converted into actual hardware drive signals. For bidirectional DC / DC circuits, the control processing module outputs PWM control signals through high-speed digital I / O ports to drive the power switching devices (such as IGBTs or MOSFETs) in the DC / DC circuit. During the switching signal generation process, a space vector modulation (SVPWM) algorithm is used to improve voltage utilization and reduce harmonic components. A dead time (typically 2 to 5 μs) is also added to prevent short circuits in the upper and lower bridge arms.

[0073] In view of the different characteristics of supercapacitors and lithium batteries, the control processing module implements differentiated control strategies: during the rapid charging and discharging process of the supercapacitor, its terminal voltage and temperature are monitored in real time, and the charging and discharging current is automatically reduced when the temperature exceeds 60°C; for lithium batteries, in addition to monitoring the voltage and temperature, the battery management system (BMS) is used to obtain the balance status of each single cell in the battery pack, and the current of each branch is dynamically adjusted during the charging and discharging process to ensure the consistency of the battery pack.

[0074] For brake resistor control, the control processing module converts the calculated PWM duty cycle into a pulse signal with a frequency of 10 to 20 kHz (the specific frequency is determined by the thermal time constant of the brake resistor). This signal drives the brake resistor's power switch via an optocoupler isolation circuit. To prevent device wear caused by frequent switching, a hysteresis control algorithm is used to set a voltage dead zone (e.g., ±5V). When the bus voltage is within this dead zone, the current duty cycle remains unchanged. The control processing module also continuously monitors the brake resistor's temperature. If the temperature exceeds 120°C, the duty cycle is automatically limited to a safe range and the air cooling system is activated.

[0075] During the entire charging and discharging process, the control processing module maintains communication with the elevator main control system through the CAN bus, uploads the energy storage system status (such as SOC, temperature, and charging and discharging power) in real time, and receives elevator operation mode switching instructions (such as maintenance mode and fire protection mode). It automatically adjusts the control strategy under special working conditions to ensure the safe coordination of the energy storage system and the overall operation of the elevator.

[0076] This embodiment adopts the above method and steps to accurately calculate energy demand and regenerated energy in future time periods by collecting elevator operation data in real time and performing dynamic power forecasting. This solves the problem of energy management lag in traditional methods, enables the energy storage system to adjust the charging and discharging strategy in advance, and improves energy recovery efficiency to more than 35% (an increase of 15 percentage points compared to traditional solutions).

[0077] The multi-parameter fusion charge and discharge priority decision-making mechanism, combined with the power change rate, supercapacitor charge state, lithium battery health status and grid electricity price period, realizes the differentiated utilization of energy storage units - supercapacitors are responsible for high-frequency energy throughput (response time <10ms), and lithium batteries handle medium and low-frequency energy storage, avoiding excessive loss of a single energy storage component, extending the life of supercapacitors by 20% and the cycle life of lithium batteries by 15%.

[0078] The dynamic adjustment technology of the braking resistor PWM duty cycle changes the traditional extensive mode of fixed duty cycle or switch quantity control. By real-time matching of regenerative energy and energy storage absorption capacity, it reduces the energy consumption of the braking resistor by 40%. At the same time, it controls the DC bus voltage fluctuation range within ±3% (the traditional solution is ±8%), significantly improving the system voltage stability.

[0079] The coordinated control strategy of the bidirectional DC / DC circuit and the braking resistor solves the problem of hardware response lag, reducing the voltage overshoot by 60% in power mutation scenarios (such as full-load emergency stop), avoiding the risk of damage to power electronic devices due to overvoltage.

[0080] The dispatching mechanism based on electricity price periods reduces the overall electricity cost of the elevator system by more than 25% through off-peak charging and peak discharging strategies, with significant economic benefits especially in high-load scenarios such as commercial buildings.

[0081] In-depth monitoring and utilization of the energy storage unit status, through the setting of SOC / SOH threshold protection strategies, effectively prevent supercapacitor overcharging (the incidence rate is reduced by 80%) and deep discharge of lithium batteries (the number of times is reduced by 30%), ensuring the long-term operation reliability of the system from the hardware level.

[0082] Overall, the method of this embodiment constructs a closed-loop control system of "data collection-predictive analysis-intelligent decision-making-precise execution", achieving technological breakthroughs in multiple dimensions such as energy recovery efficiency, voltage stability, equipment life protection, and electricity cost optimization. It comprehensively improves the overall performance of the elevator energy storage system and provides an efficient and reliable solution for the engineering application of elevator energy-saving technology.

[0083] The charge and discharge control method for an elevator energy storage system of this embodiment may further specifically include the following steps: S301, acquiring real-time operation data of the elevator.

[0084] This step can refer to the description of the above embodiment and will not be repeated here.

[0085] S302: Determine the instantaneous generated power and power change rate of the elevator traction machine based on the real-time operating data.

[0086] This step can refer to the description of the above embodiment and will not be repeated here.

[0087] S303: Generate a multi-dimensional feature data set based on the real-time operation data and the historical power change rate database.

[0088] Specifically, the control processing module first collects real-time elevator operating data through its interface. While acquiring this data, the control processing module accesses an internally stored or connected historical power rate change database to retrieve historical power rate change data. This historical data is not simply recorded; it is organized and annotated, and associated with specific elevator operating conditions (such as similar positions, loads, and speed combinations).

[0089] Next, the control processing module integrates and processes real-time data with relevant historical data. It extracts key features, which may include but are not limited to current power, acceleration, historical average power rate of change, historical peak power rate of change under specific operating conditions, current operating direction, and distance to the target floor. These raw data and derived features undergo cleaning (to remove outliers or noise), normalization (to eliminate dimensional effects between different features and keep them within a comparable range), and necessary format conversion to ultimately construct a structured, multi-dimensional feature dataset.

[0090] Each row of this multi-dimensional feature dataset represents a specific time point or system state, and each column represents a feature dimension that affects the predicted power, ensuring data integrity and consistency.

[0091] In this embodiment, the construction process of the historical power change rate database includes: during the long-term operation of the elevator, continuously collecting real-time operating data such as car position, load, traction motor speed, DC bus voltage and current under different operating conditions (such as no load, full load, start, braking, and stopping at different floors), calculating the instantaneous generated power at each sampling moment and the power change rate of adjacent moments (that is, the power change per unit time) based on the above data, and synchronously recording the corresponding elevator operating status parameters (such as running direction, speed, number of floors), energy storage unit status (supercapacitor charge state, lithium battery health status) and grid electricity price period and other related information. After data cleaning (eliminating outliers and filling missing values) and normalization, the data are classified and stored in a structured database according to time series or operating condition type, forming a data set containing historical power change rates and multi-dimensional related features.

[0092] S304: Input the multi-dimensional feature data set into a pre-trained XGBoost regression model to obtain an initial power change prediction curve including power prediction values ​​at different time points within a future preset time period.

[0093] This data is then fed into a pre-trained XGBoost regression model. This XGBoost model is trained using offline or online machine learning methods based on a large amount of historical operating data and corresponding power results. It has already learned and solidified the complex nonlinear relationship between elevator power changes and the aforementioned multi-dimensional features. The control processing module encapsulates the prepared feature dataset in the specific format required by the model, ensuring that each feature corresponds to the input dimension used during model training. This module then uses these as input parameters to invoke and initiate the model's internal prediction process.

[0094] Within the model, XGBoost (Extreme Gradient Boosting) utilizes its core gradient boosting decision tree ensemble algorithm for efficient computation. The input feature vector is passed and evaluated along a path through a series of decision trees integrated within the model. Each tree contributes a predicted value to the final power value based on its learned splitting rules. The model then accurately weights and aggregates the predictions of all these base learners (decision trees) to produce one or a series of highly accurate predictions.

[0095] Specifically, the control processing module sets a preset time period in the future (for example, the next 5 or 10 seconds) and instructs the model to output power predictions at multiple discrete time points within this period. These predictions are then linked together to form an initial curve that reflects the elevator's power variation trend in the short term. This initial power variation prediction curve provides fundamental predictive information for subsequent energy storage decisions.

[0096] In this embodiment, the training process of the XGBoost regression model includes: First, a multi-dimensional feature dataset is extracted from the historical power change rate database, including real-time elevator operation parameters (such as car position, load capacity, and traction motor speed), historical power change rate, energy storage unit status (supercapacitor charge state, lithium battery health status), and grid electricity price period. The dataset is divided into training set, validation set, and test set in an 8:1:1 ratio.

[0097] Feature engineering was performed on the training set, including missing value processing (such as mean filling), categorical feature encoding (such as converting electricity price periods into numerical values), and normalization and scaling. The XGBoost regression model was initialized, and the objective function was set to the mean squared error (MSE) to minimize the difference between the predicted power and the actual power. The gradient boosting framework was used for iterative training, and a new decision tree was generated in each iteration to fit the residual of the previous round.

[0098] During training, an early stopping strategy is implemented on the validation set (e.g., stopping if the validation error does not improve after 10 consecutive rounds) to prevent overfitting, and regularization parameters (e.g., gamma, lambda) are used to control model complexity. Cross-validation (e.g., 5-fold cross-validation) is used to evaluate the generalization ability of the model, and hyperparameters such as the learning rate, tree depth, and subsampling rate are dynamically adjusted to optimize performance.

[0099] After training is completed, the test set is used to evaluate the final performance indicators of the model (such as root mean square error RMSE, mean absolute error MAE, determination coefficient R 2 ), and determine the contribution of each input feature to power prediction through feature importance analysis, and finally obtain an XGBoost regression model that can be used to predict the power changes of the elevator traction machine within a preset time period in the future.

[0100] S305 , based on the real-time detected traction machine vibration spectrum data, error compensation is performed on the initial power prediction curve through a sliding time window to obtain a power change prediction value of the elevator traction machine.

[0101] Specifically, first, vibration sensors (such as accelerometers) installed at key parts of the traction machine are used to collect vibration signals in real time, and the sampling frequency is set to 10kHz to cover the characteristic frequency of the traction machine fault (usually below 1kHz). The collected time-domain vibration signal is subjected to a fast Fourier transform (FFT) to convert it into a frequency-domain signal to obtain a vibration spectrum containing each frequency component and its amplitude. At the same time, time series data is extracted from the initial power prediction curve output by the XGBoost regression model, and the sliding time window is divided into fixed lengths (such as 10 sampling points). In each time window, the control processing module calculates the correlation coefficient between the characteristic parameters of the vibration spectrum (such as the main frequency amplitude, frequency band energy distribution) and the power prediction value, and establishes a mapping relationship between the two.

[0102] For example, when a sudden increase in energy in the 200Hz-300Hz frequency band is detected, it indicates that the traction machine may enter a braking state. At this time, the power prediction value at the corresponding time point will be corrected. The control processing module uses the Kalman filter algorithm to fuse the vibration spectrum characteristics and the power prediction value: the vibration spectrum characteristics are used as external input, the power prediction value is used as the system state, and the prediction value is iteratively updated through the state transfer equation and the observation equation. To cope with sudden working conditions, when the rate of change of the characteristic parameters of the vibration spectrum exceeds the preset threshold (such as the rate of change of the main frequency amplitude >30%), the control processing module activates the adaptive weight adjustment mechanism to increase the weight of the vibration data in error compensation.

[0103] The control processing module corrects the initial power prediction curve point by point through a sliding time window and outputs the compensated power change prediction value, so that the prediction result is closer to the actual operating status of the traction machine.

[0104] S306: Reset the positive values ​​in the power change prediction value to zero and retain the negative values ​​to obtain a generated power sequence.

[0105] Specifically, time series data is first extracted from the aforementioned power change predictions. Each data point contains a prediction timestamp and the corresponding power prediction value. The control processing module traverses this time series and performs a sign check on each power prediction value: if the power prediction value is positive, it indicates that the elevator is consuming power (such as during upward movement or acceleration), and the value is set to zero; if the power prediction value is negative, it indicates that the elevator is generating power (such as during downward movement or braking), and the negative value is retained. This process forms a sequence containing only power data in the power generation state, namely the power generation power sequence.

[0106] To ensure data continuity, the control processing module retains the original timestamp information during the zeroing operation, aligning the generated power series with the original forecast curve in the temporal dimension. The control processing module then smoothes the generated power series using a moving average filter (e.g., with a window size of 3) to eliminate any high-frequency noise while maintaining the trend characteristics of the series.

[0107] To verify the validity of the processing results, the control processing module calculates the cumulative energy value of the generated power sequence and compares it with the theoretical regenerative energy estimated by the elevator physical model. If the deviation exceeds 15%, the data backtracking mechanism is triggered to check whether there is any abnormality in the vibration spectrum compensation in step S305.

[0108] The generated power sequence that has undergone symbol screening, smoothing, and validity verification will be used in subsequent integration operations to accurately calculate the total amount of generated energy that the elevator energy storage system needs to store within a preset time period in the future.

[0109] S307 , performing an integration operation on the generated power sequence within the future preset time period to obtain a theoretical storable amount of electricity.

[0110] Specifically, the control processing module first extracts discrete power data points from the power sequence within a preset future time period (e.g., 10 seconds). Each data point includes a timestamp and the corresponding power value. The module then divides this time period into equally spaced time slices (e.g., 100ms) and numerically integrates the power sequence using the trapezoidal integration method.

[0111] For two adjacent data points P(t_i) and P(t_{i+1}), the control processing module calculates the time interval Δt = t_{i+1} - t_i and calculates the energy element within that time slice using the formula E_i = [(P(t_i) + P(t_{i+1})) / 2] × Δt. The control processing module traverses all adjacent pairs of points in the power generation sequence, accumulates the energy elements of each time slice, and obtains a preliminary estimate of the theoretically storable power.

[0112] To improve integration accuracy, when the time interval between adjacent data points is greater than a preset threshold (e.g., 200ms), linear interpolation can be automatically performed to insert additional calculation points between the two points. The continuity of the generated power sequence is also checked. If missing data is found (e.g., no power value in a certain time slice), cubic spline interpolation is used to complete the data.

[0113] After the integration operation is complete, the result is converted to a standard energy unit (such as joules or kilowatt-hours) and compared with the theoretical value calculated by the elevator physical model. If the deviation exceeds 5%, the validity of the power generation sequence and the integration parameter settings are rechecked. Finally, the control processing module outputs the theoretical storable power capacity after optimizing the accuracy and verifying the validity.

[0114] S308 , performing loss compensation on the theoretically storable amount of electricity according to the real-time charging and discharging efficiency of the energy storage unit in the elevator energy storage system, to obtain the total amount of generated energy that actually needs to be stored.

[0115] Specifically, the charge and discharge efficiency parameters of the energy storage unit (supercapacitor and lithium battery) are first obtained in real time through the CAN bus. For supercapacitors, their current temperature and state of charge (SOC) are read, and the preset efficiency mapping table is queried to obtain the corresponding charging efficiency η_SC (usually 0.92-0.98) and discharge efficiency η_DC (usually 0.95-0.99). For lithium batteries, in addition to considering temperature and SOC, the health status (SOH) parameters are also combined to calculate their charge and discharge efficiency through the formula η_Batt = η_base × (1-k × (100% - SOH)), where η_base is the baseline efficiency (usually 0.90-0.95) and k is the attenuation coefficient (usually 0.005-0.01).

[0116] The control processing module constructs an energy loss compensation model based on the theoretical storable amount of energy E_theory obtained in step S307 and the real-time charge and discharge efficiency of the energy storage unit. If the energy storage unit is currently charging, the actual amount of generated energy to be stored is calculated as E_actual = E_theory / (η_SC × η_Batt), taking into account the dual energy losses during the charging process.

[0117] If the energy storage unit is in a discharging state, the module compensates for losses using E_actual = E_theory × (η_DC × η_Batt). To address dynamic changes in efficiency parameters, a sliding time window (e.g., the last five sampling cycles) is used to track the changing trends of charge and discharge efficiency. If the efficiency fluctuation exceeds 3%, the adaptive adjustment mechanism is triggered and the compensation coefficient is recalculated.

[0118] Furthermore, the final result can be corrected based on the grid feedback efficiency (typically 0.90-0.93) to ensure that the total energy calculation covers all energy conversion paths. Ultimately, the actual total amount of stored generated energy, after dynamic compensation for charge and discharge efficiency, is output, providing a precise energy target value for the energy storage system's charge and discharge control.

[0119] S309: Determine a fluctuation state of the current load of the elevator traction machine based on the power change rate, where the fluctuation state includes high-frequency transient fluctuation or steady-state fluctuation.

[0120] Specifically, continuous power change rate data points are first extracted from the real-time operating data to form a time series. The control processing module calculates the standard deviation σ and average change rate μ of this series as fluctuation characteristic indicators. If σ exceeds a preset high-frequency threshold (such as 1.5kW / ms) and the absolute value of μ is greater than a steady-state threshold (such as 0.8kW / ms), it is determined to be a high-frequency transient fluctuation. If σ is lower than a low-frequency threshold (such as 0.5kW / ms) and μ fluctuates within a small range around zero (such as ±0.3kW / ms), it is determined to be a steady-state fluctuation.

[0121] To improve the accuracy of the determination, the control processing module uses a sliding window (e.g., 500ms) for time-domain analysis and calculates the kurtosis K of the power change rate within the window. When K > 3, the data distribution has a peaked, thick-tailed characteristic, corresponding to a high-frequency shock load. When K ≈ 3, the data approximates a normal distribution, corresponding to a steady load.

[0122] The control processing module also makes auxiliary judgments based on the elevator operation stage: when the car is in the starting acceleration or braking deceleration stage, if the spectrum analysis of the power change rate shows that the frequency components above 10Hz account for more than 60%, the judgment of high-frequency transient fluctuations is strengthened; when the car is running at a constant speed, if the power change rate remains within the range of ±0.5kW / ms for a long time, it is confirmed as a steady-state fluctuation.

[0123] The control processing module compares the current fluctuation state with the historical pattern library, further verifies the judgment result through the pattern recognition algorithm, and finally outputs a description of the load fluctuation state including the fluctuation type, frequency range, and impact intensity.

[0124] S310: Determine the availability of an energy storage unit according to the state of charge of the supercapacitor and the health state of the lithium battery.

[0125] Specifically, the control processing module first obtains the real-time state of charge (SOC_SC) of the supercapacitor and the state of health (SOH_Batt) of the lithium battery through the CAN bus.

[0126] For supercapacitors, if the SOC_SC is lower than 20%, it is marked as "low availability" to limit its further discharge; if the SOC_SC is higher than 85%, it is marked as "high availability" and is used first to absorb regenerative energy; if the SOC_SC is in the range of 20%-85%, the availability weight is dynamically adjusted according to the power change rate.

[0127] For lithium batteries, when SOH_Batt is lower than 50%, it is marked as "limited availability" and is only allowed to charge and discharge within a safe current range; when SOH_Batt is higher than 80%, it is marked as "fully available" and can undertake the main energy storage task.

[0128] The control processing module calculates the remaining available capacity of the supercapacitor C_SC=C_SC_max×(1-SOC_SC) and the equivalent available capacity of the lithium battery C_Batt=C_Batt_max×SOH_Batt×SOC_Batt, where C_SC_max and C_Batt_max are the rated capacities respectively.

[0129] If C_SC+C_Batt is less than 30% of the total system demand, the overall availability of the energy storage unit is determined to be insufficient, triggering the grid feedback priority enhancement mechanism.

[0130] The control processing module also considers the charge and discharge rate limit: when the power change rate exceeds the maximum response rate of the supercapacitor (such as 10kW / ms), its priority will be reduced even if the SOC_SC is in the high availability range; when the temperature of the lithium battery exceeds 55°C, its charge and discharge current upper limit will be automatically lowered and the availability rating will be adjusted accordingly.

[0131] The control processing module integrates the above factors and generates an energy storage unit availability index (0-100 points) through a fuzzy logic algorithm, and converts it into decision labels such as "priority use", "cautious use", and "restricted use", providing a quantitative basis for determining charging and discharging priorities.

[0132] S311 , determining a current electricity price period based on a real-time clock and an electricity price database, wherein the electricity price period includes a peak electricity period, an off-peak electricity period, or a normal electricity period.

[0133] Specifically, the system first uses the real-time clock to obtain the current date and time, accurate to the minute. The control processing module then extracts the region's time-of-use electricity pricing policy from a locally stored electricity price database. This database contains the rules for dividing electricity prices into peak, off-peak, and normal periods for 365 days a year. For example, peak periods are 8:00-11:00 and 18:00-23:00, off-peak periods are 23:00-7:00 the following day, and the rest are normal periods.

[0134] The control processing module compares the current time with the time period boundaries in the database to determine the current electricity price period. If a special date (such as a holiday) occurs, the control processing module automatically calls the special time period rules in the database to make adjustments. To accommodate dynamic updates to electricity pricing policies, the control processing module regularly downloads the latest electricity price data from the power company's server via a network interface (e.g., weekly) and automatically updates the local database. If the network connection is interrupted, the control processing module uses the most recently downloaded valid data and records an exception log.

[0135] The control processing module also calculates the remaining duration of the current period, providing a time-based reference for subsequent charging and discharging strategies. For example, if the current peak power period is less than an hour, the control processing module will prioritize energy storage units, reducing grid power draw. The control processing module outputs comprehensive information including the price period type, current price, and remaining time in the period, serving as a key input for charging and discharging priority decisions.

[0136] S312: Input the fluctuation state, the availability of the energy storage unit, and the current electricity price period into a preset rule library to obtain the charging and discharging priority of the supercapacitor, lithium battery, or grid feedback of the elevator energy storage system.

[0137] Specifically, the control processing module first normalizes the aforementioned fluctuation state (high-frequency transient fluctuation or steady-state fluctuation), energy storage unit availability (quantified as an index ranging from 0 to 100), and electricity price period information. The control processing module then maps these input parameters to corresponding rule groups in a pre-defined rule base, which uses production rules and contains over 200 decision rules.

[0138] For example, when the fluctuation state is high-frequency transient fluctuation, the supercapacitor availability index is greater than 80 and it is in the peak power period, the rule of "supercapacitor priority charging > lithium battery charging > grid feedback" is triggered; when the fluctuation state is steady-state fluctuation, the lithium battery availability index is greater than 70 and it is in the valley power period, the rule of "lithium battery priority discharge > supercapacitor discharge > grid power extraction" is triggered.

[0139] The control processing module traverses the rule base using a forward inference mechanism, calculates the matching degree of each rule's prerequisites, and comprehensively evaluates the confidence level of each rule using a weighted summation method. For conflicting rules (e.g., multiple rules that simultaneously meet some conditions), the control processing module resolves the conflict in a hierarchical order, with electricity price period being the highest priority, followed by fluctuation status, and finally energy storage unit availability.

[0140] The control processing module also considers device safety constraints. For example, when the supercapacitor temperature exceeds 65°C, its charging and discharging priority is automatically reduced, even if other conditions are met.

[0141] Finally, the control processing module outputs a decision result including the charging and discharging priority order (such as "supercapacitor > lithium battery > grid feedback") and its confidence level. This result will directly drive the mode switching of the bidirectional DC / DC circuit and the PWM duty cycle adjustment of the braking resistor.

[0142] S313: Generate a switching signal for a charge and discharge mode according to the charge and discharge priority, where the charge and discharge mode includes a supercapacitor priority charging mode, a lithium battery priority discharging mode, or a grid feedback mode.

[0143] Specifically, the control processing module first analyzes the charge and discharge priority order output in step S312 and converts it into the corresponding charge and discharge mode. If the priority is "supercapacitor > lithium battery > grid feedback", a "supercapacitor priority charging mode" switching signal is generated; if the priority is "lithium battery > supercapacitor > grid power", a "lithium battery priority discharging mode" switching signal is generated; if the priority is "grid feedback > supercapacitor > lithium battery", a "grid feedback mode" switching signal is generated.

[0144] The control processing module assigns a unique digital code to each charging and discharging mode (e.g., 001 represents supercapacitor priority charging mode) and encapsulates it into a CAN bus message frame containing the mode type, effective time, and priority level. To ensure the reliability of the switching signal, the control processing module verifies the message content using the CRC-16 checksum algorithm and regenerates the message if the check fails.

[0145] The control processing module will also perform signal correction according to the current status of the energy storage unit. For example, when the supercapacitor SOC exceeds 90%, even if the priority is supercapacitor charging, it will automatically switch to lithium battery charging mode and generate a corrected switching signal.

[0146] Furthermore, the control processing module records historical data for each mode switch, including the switch time, triggering reason, and previous and subsequent mode types, for subsequent system performance analysis. Ultimately, the control processing module outputs an encoded and verified charge / discharge mode switch signal through the hardware I / O port and stores a copy of the signal in a local log file.

[0147] S314 , controlling the bidirectional DC / DC circuit to switch the charge and discharge mode based on the switching signal.

[0148] Specifically, the control processing module first receives the switching signal generated in step S313 and obtains the charge and discharge mode code by parsing the CAN message frame. The control processing module queries the predefined mode-control parameter mapping table to obtain the bidirectional DC / DC circuit control parameters corresponding to the mode, including the switching frequency (usually 20kHz-50kHz), duty cycle range, current limit, etc. For example, in the supercapacitor priority charging mode, the control processing module sets the DC / DC circuit to buck mode, the switching frequency to 40kHz, the initial duty cycle to 0.6, and the charging current upper limit to 150A.

[0149] The control processing module sends configuration commands to the DC / DC circuit's digital controller via the SPI bus. These commands contain parameters such as the frequency setting register value, the duty cycle register value, and the protection threshold. After sending the command, the control processing module activates a timeout monitoring mechanism (set to 5ms) and waits for the DC / DC circuit to return an acknowledgment message. If no acknowledgment is received within the timeout, the control processing module retries the command up to three times. If this fails, fault handling procedures are triggered, such as switching to safe mode and generating an alarm.

[0150] The control processing module uses a soft-start strategy during mode switching, smoothly transitioning control parameters from their current values ​​to their target values ​​within 200ms through linear interpolation, preventing sudden changes in current and voltage from impacting the circuit. The control processing module monitors the DC / DC circuit's operating status in real time, including parameters such as input and output voltage, current, and temperature. If an anomaly (such as overvoltage or overcurrent) is detected, the protection mechanism is immediately triggered, adjusting control parameters or interrupting the switching process.

[0151] In this way, the control processing module completes the charging and discharging mode switching of the bidirectional DC / DC circuit, confirms the success of the switching through the state feedback mechanism, and records the switching result in the system log.

[0152] S315 , after switching the charge and discharge mode, determining the brake resistor PWM duty cycle according to the detected real-time deviation of the bus voltage.

[0153] Specifically, the control processing module first samples the DC bus voltage at a sampling frequency of 10kHz. After converting it to a digital signal using a 16-bit ADC, it uses a sliding average filter (for example, with a window size of 5) for denoising to obtain the real-time bus voltage value V_actual. Simultaneously, the target bus voltage V_target corresponding to the current charge / discharge mode (for example, 750V in supercapacitor charging mode) is retrieved from the system parameter table, and the deviation between the two is calculated as ΔV = V_actual - V_target.

[0154] The control processing module inputs ΔV into a fuzzy-PID-based controller, which consists of three submodules: a fuzzification module maps ΔV and the rate of change of ΔV d(ΔV) / dt to linguistic variables (such as "negative large," "zero," and "positive small"); a rule reasoning module performs logical reasoning based on 49 preset fuzzy rules (such as "if ΔV is positive and large and d(ΔV) / dt is positive and small, then the output duty cycle increment is positive and small"); and a defuzzification module uses the center of gravity method to convert the reasoning result into a precise duty cycle increment ΔD.

[0155] To prevent the brake resistor from overheating, the control processing module monitors its temperature T in real time, obtains the duty cycle correction coefficient k_T corresponding to the current temperature through a table lookup (e.g., k_T = 0.9 when T = 80°C), and calculates the duty cycle increment ΔD' = ΔD × k_T after temperature constraint. The control processing module further introduces a historical action frequency constraint: it counts the number of brake resistor actions N in the past 10 seconds. When N > 5, it calculates the frequency correction coefficient using the formula k_N = 1 - 0.05 × (N - 5), resulting in the final duty cycle increment ΔD" = ΔD' × k_N. The control processing module adds the current duty cycle D_current to ΔD" to obtain the new duty cycle D_new, and constrains it to the safe range [0.1, 0.9] using a limiter function.

[0156] To avoid frequent adjustments, the control processing module sets a voltage dead zone of ±5V, and keeps the current duty cycle unchanged when |ΔV|≤5V.

[0157] The control processing module can generate a PWM signal at a set frequency (e.g., 20kHz), and control the power MOSFET of the braking resistor through an optocoupler isolation drive circuit to achieve precise regulation of the bus voltage.

[0158] During the entire process, the control processing module continuously records parameters such as bus voltage, duty cycle, and brake resistor temperature, and updates the control decision every 50ms to form a closed-loop control loop to ensure that the bus voltage fluctuation does not exceed ±3% under various operating conditions.

[0159] In some embodiments, step S315 may include steps S3151-S3154, specifically as follows: S3151: Construct a sliding mode surface in a sliding mode control algorithm based on the real-time deviation and change rate of the bus voltage.

[0160] The sliding mode control algorithm is a robust control strategy based on variable structure systems. By designing a specific sliding surface, the system state trajectory converges to this hyperplane within a finite time, and then slides toward the equilibrium point on the sliding surface with preset dynamic characteristics. Its core lies in constructing a sliding mode function (such as s = λe + de / dt) that contains the system error and its derivative. When the system state reaches the sliding mode surface, a switching control law (such as a sign function or a saturation function) generates high-frequency switching actions, forcing the system to move along the sliding surface. This strategy is insensitive to parameter perturbations and external disturbances.

[0161] In this embodiment, the sliding mode control algorithm constructs a sliding mode surface by calculating the bus voltage deviation and its rate of change in real time. When a state deviation is detected, the PWM duty cycle is adjusted at a μs-level response speed to achieve fast and accurate regulation of the braking resistor power.

[0162] When executing step S3151 , the control processing module may first obtain the real-time value V_actual of the DC bus voltage at a sampling frequency of 10 kHz, and calculate the deviation e=V_actual−V_target from the real-time value V_actual of the DC bus voltage.

[0163] At the same time, the rate of change of the voltage deviation is calculated by a differential algorithm: de / dt=(e-e_prev) / T_s, where e_prev is the voltage deviation of the previous sampling period, and T_s is the sampling period (0.1 ms).

[0164] The control processing module constructs a sliding surface function s=e+λ·∫e·dt, where λ is a sliding surface parameter (the value can be 5), and ∫e·dt is calculated in real time by the trapezoidal integration method.

[0165] To improve the robustness of the system, the control processing module uses an exponential reaching law to design a sliding mode controller. The reaching law expression is ds / dt = -k·sat(s / φ), where k is the gain coefficient (set to 10), φ is the boundary layer thickness (set to 0.5), and sat is the saturation function.

[0166] The control processing module maps the voltage deviation e and rate of change de / dt into state space and dynamically updates the sliding surface position on the state plane, ensuring that the system state point always moves on the sliding surface. The control processing module uses a Kalman filter to estimate the state of e and de / dt, filtering out the effects of measurement noise on the sliding surface construction.

[0167] The control processing module also dynamically adjusts the sliding surface parameter λ according to the current charge and discharge mode: in the supercapacitor fast charging mode, λ is increased to 8 to enhance the system response speed; in the lithium battery constant current discharge mode, λ is reduced to 3 to reduce system vibration.

[0168] The control processing module constructs a sliding mode surface that can reflect the dynamic characteristics of the bus voltage, providing a basis for subsequent sliding mode control calculations.

[0169] In some embodiments, the functional expression of the sliding surface may include: Among them, ΔV represents the real-time deviation of bus voltage; represents the fractional differential term of ΔV; α is the order of the fractional differential operator; t is time; λ(t) represents the time-varying coupling coefficient; λ0 is the initial coupling coefficient, which controls the reference weight of the fractional differential term in the sliding surface; sign(ΔV) represents the sign function based on ΔV. It outputs +1 when the bus voltage deviation ΔV>0 and -1 when ΔV<0. It is used to quickly guide the voltage deviation toward zero in sliding mode control. Its essence is to provide a nonlinear control direction switching signal. γ is the potential field strength coefficient, which controls the contribution of the nonlinear potential field term to the convergence characteristics of the sliding surface; β is the potential field attenuation coefficient, which adjusts the attenuation rate of the exponential term as the voltage deviation increases; τ represents the integral time variable, which indicates the cumulative time process from the initial moment to the current moment; k is the cumulative gain coefficient, which determines the growth rate of the voltage deviation history integral to the time-varying coupling coefficient λ(t); e is a natural constant.

[0170] The sliding surface function expression of this embodiment is analyzed from the following aspects.

[0171] 1. Multi-dimensional state description to enhance dynamic adaptability: Fractional differentials Breaking through the limitations of traditional integer-order differentials, this approach uses fractional-order operators with a range of 0<α<1 to accurately characterize the "memory-heritage" characteristics of bus voltage deviations. In elevator braking scenarios, the intermittent injection of motor regenerative energy (such as during car start-up, stop-down, acceleration, and deceleration) causes non-stationary fluctuations in bus voltage. Fractional-order differentials can capture more subtle dynamic changes (such as the gradual change trend after a sudden voltage change). Compared to integer-order differentials (which only reflect the instantaneous rate of change), they can more comprehensively describe system states and improve the adaptability of sliding mode systems to complex operating conditions.

[0172] Time-varying coupling coefficient λ(t): The amplitude of the voltage deviation accumulated by integration Dynamically adjust the weight of the differential term. For example, when an elevator stops suddenly (bus voltage suddenly rises), λ(t) increases rapidly, strengthening differential feedback and accelerating the system's approach to the sliding mode surface. During steady-state operation (with small deviations), λ(t) maintains a low gain to avoid overshoot. This time-varying characteristic enables the sliding mode surface to adapt to scenarios such as sudden load changes and energy recovery intensity variations, resolving the response lag problem of traditional fixed-coefficient sliding mode surfaces under strong disturbances.

[0173] 2. Chattering suppression and energy optimization: Potential field correction term (γ·sign(ΔV)·e -β|ΔV| ): Introducing an exponentially decaying potential field function to achieve "quick response to strong deviations and smooth adjustment to weak deviations". When |ΔV| is large (such as in the initial stage of braking), e -β|ΔV| ≈1, the potential field term provides strong feedback force, rapidly reducing voltage deviation. As the deviation approaches zero (e.g., near the target voltage of 750V), the potential field term decays exponentially, reducing high-frequency switching (jittering) of the control input. Compared to traditional sign functions (which cause high-frequency jitter and increase switching losses in the brake resistor), this design, through the continuously decaying potential field force, reduces the switching frequency of the PWM signal while maintaining regulation accuracy (within ±5V), extending the life of power devices (such as MOSFETs), and meeting the energy efficiency and reliability requirements of elevator systems.

[0174] 3. Engineering feasibility and scenario matching: Computational complexity adaptation: Fractional-order differentiation can be approximated using digital algorithms such as the Oustaloup filter. (The DSP chip in the control processing module supports floating-point operations and can complete the calculation within a 50μs period, meeting a 20kHz control frequency.) The integral operation (trapezoidal method) and exponential operation (table lookup or Taylor expansion) of the time-varying coupling coefficient are both low-complexity operations, ensuring real-time performance.

[0175] Elevator braking scenario adaptation: The elevator bus voltage needs to be stable within milliseconds (for example, GB / T 10058-2009 requires voltage fluctuations ≤±7%). The sliding surface is corrected in multiple dimensions (deviation, differential, time-varying, potential field) to control the voltage deviation within ±5V (corresponding to 0.67% fluctuation of a 750V bus) under a wide range of operating conditions (light load / heavy load, 1m / s to 6m / s speed braking), which is far better than the standard requirements. At the same time, the sign function sign (ΔV) of the potential field term can be adjusted in both directions (applicable to both boost / buck scenarios), covering all operating conditions of elevator regenerative braking (bus voltage increases, energy needs to be discharged) and grid power extraction (bus voltage decreases, energy needs to be replenished). 4. Theoretical innovation and robustness improvement: Fusion of fractional order and potential field theory: Fractional order calculus (for handling complex dynamics) is combined with potential field method (for suppressing chattering) to construct a sliding mode surface with both high dynamic response and low chattering, breaking through the limitations of traditional sliding mode control in nonlinear systems. In the strongly coupled and time-varying scenarios of elevator braking, compared with traditional sliding mode surfaces (such as , the robustness to model uncertainties (such as motor parameter drift and load mutation) is improved by more than 30% (verified by simulation: under ±20% load disturbance, the voltage deviation is still ≤±5V, while the traditional method deviation is as high as ±15V).

[0176] This embodiment adopts the above-mentioned sliding surface function. Through mathematical multi-dimensional state description, real-time feasibility in engineering, and deep adaptation to elevator braking scenarios, it is reasonable both in theory and practice, and provides an effective solution for high-precision, low-jitter control of bus voltage.

[0177] S3152: When the absolute value of the sliding mode surface is greater than the sliding mode threshold, a sliding mode control calculation is performed to obtain a first duty cycle increment.

[0178] Specifically, the absolute value |s| of the sliding surface s constructed in step S3151 is first calculated and compared with the preset sliding mode threshold (set to 0.8). When |s|>0.8, it indicates that the system state point deviates from the sliding mode surface, and the control processing module starts the sliding mode control calculation. According to the exponential convergence law ds / dt=-k·sat(s / φ), the control processing module calculates the control input u=u_eq+u_s, where u_eq is the equivalent control term, and u_eq=-λ·e is obtained by setting ds / dt=0; u_s is the switching control term, u_s=-k·sat(s / φ). The control processing module converts the control input u into a duty cycle increment ΔD_1, and the conversion formula is ΔD_1=k_p·u, where k_p is the proportional coefficient (set to 0.02).

[0179] To suppress high-frequency chattering, the control processing module replaces the traditional sign function with a saturation function: when |s / φ| ≤ 1, sat(s / φ) = s / φ; when |s / φ| > 1, sat(s / φ) = sign(s). The control processing module also implements an anti-integral windup mechanism that suspends updates to the integral term when the accumulated duty cycle increment exceeds 0.2.

[0180] The control processing module also dynamically adjusts the gain factor k based on the bus voltage trend: when |de / dt| > 2V / ms, k is increased to 15 to accelerate system response; when |de / dt| < 0.5V / ms, k is reduced to 8 to reduce chattering. The control processing module performs sliding-mode control calculations every 0.1ms, updates the duty cycle increment ΔD_1, and stores the results in a ring buffer for subsequent gradient constraint processing.

[0181] The control processing module obtains the first duty cycle increment through a sliding mode control algorithm to achieve fast tracking and robust control of the bus voltage.

[0182] S3153: Perform a gradient constraint on the first duty cycle increment according to the real-time temperature of the braking resistor to obtain a second duty cycle increment.

[0183] Specifically, a temperature sensor (such as an NTC thermistor) mounted on the surface of the brake resistor collects the temperature T in real time at a sampling frequency of 100Hz. After signal conditioning (amplification and filtering) and ADC conversion, the digital temperature value is obtained. The control processing module then consults a preset temperature-gradient constraint mapping table, which defines the duty cycle increment adjustment coefficient k_T for different temperature ranges based on the material properties and heat dissipation capacity of the brake resistor.

[0184] For example, when the temperature T<60℃, k_T=1.0 (no constraint); when 60℃≤T<80℃, k_T=0.9 (gradient reduced by 10%); when 80℃≤T<100℃, k_T=0.7 (gradient reduced by 30%); when T≥100℃, k_T=0.5 (forced reduction of 50%), and a temperature warning is triggered.

[0185] The control processing module multiplies the first duty cycle increment ΔD_1 obtained in step S3152 by k_T to obtain a second duty cycle increment ΔD_2 = ΔD_1 × k_T subject to temperature constraints.

[0186] To avoid regulation oscillations caused by rapid temperature changes, the module uses a first-order inertial filter algorithm to smooth the temperature signal, with a time constant set to 0.5 seconds. If the temperature continues to rise within 10 seconds and the cumulative decrease in k_T exceeds 40%, the control processing module automatically limits the absolute value of ΔD_2 to no more than 0.1 to prevent the brake resistor from overheating and failure.

[0187] The module uses a temperature gradient constraint mechanism to control the brake resistor temperature within a safe threshold (120°C) while ensuring the stability of the bus voltage.

[0188] S3154: Generate a PWM duty cycle of the braking resistor based on the second duty cycle increment, the real-time temperature of the braking resistor, and the historical operation frequency.

[0189] Specifically, first, maintain a historical action frequency queue with a preset length (e.g., 10 seconds) to record the switching moments of the PWM signals of the braking resistor. By calculating the number of actions N in the queue (defined as the number of times the voltage deviation exceeds the dead zone and the duty cycle changes), generate a frequency correction coefficient k_N. When N ≤ 3, k_N = 1.0 (no correction); when 3 < N ≤ 6, k_N = 0.9 (the higher the frequency, the lower the correction coefficient); when N > 6, k_N = 0.8 (forcefully reduce the correction coefficient to reduce the action frequency).

[0190] The control processing module multiplies the second duty cycle increment ΔD_2 by k_N to obtain ΔD_3 = ΔD_2 × k_N, and then adds it to the current duty cycle D_current to obtain the preliminary duty cycle D_pre = D_current + ΔD_3.

[0191] To ensure safety, constrain D_pre within the range of [0.05, 0.95] through a limiting function (to avoid full - on or full - off states), and make a judgment in combination with the bus voltage dead zone (±5V): If the absolute value of the current voltage deviation ≤ 5V, then D_new = D_current (maintain the current duty cycle); otherwise, D_new = D_pre.

[0192] The control processing module can further adjust the final duty cycle according to the real - time temperature T of the braking resistor: When T ≥ 110°C, D_new = D_new × 0.9 (forcefully derate); when T < 50°C, D_new = D_new × 1.1 (allow a moderate increase).

[0193] Finally, the control processing module generates PWM signals at a set frequency (e.g., 20kHz), controls the power switch of the braking resistor through an opto - isolator drive circuit, and stores data such as the current duty cycle, temperature, and frequency in a circular buffer for subsequent dynamic optimization of the control strategy, forming a closed - loop control link of "detection - calculation - constraint - output".

[0194] S316, based on the charge - discharge mode and the PWM duty cycle of the braking resistor, control the elevator energy storage system to perform charge - discharge.

[0195] This step can refer to the description of the foregoing embodiments and will not be elaborated here.

[0196] The method provided in the above embodiments can be executed by an electronic device. The following describes this electronic device in the embodiments of the present invention from the perspective of hardware processing. Please refer to Figure 3 which is a schematic structural diagram of a physical device of the electronic device in the embodiments of the present invention.

[0197] It should be noted that Figure 3The structure of the electronic device shown is only an example and should not limit the functions and scope of use of the embodiments of the present invention.

[0198] like Figure 3 As shown, the electronic device includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the storage part 408 into the random access memory (RAM) 403, such as the method described in the above embodiment. In the random access memory (RAM) 403, various programs and data required for system operation are also stored. The central processing unit (CPU) 401, the read-only memory (ROM) 402 and the random access memory (RAM) 403 are connected to each other via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0199] The following components are connected to the input / output (I / O) interface 405: an input section 406 including an audio input device, a push button switch, etc.; an output section 407 including a display, an audio output device, an indicator light, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the input / output (I / O) interface 405 as needed. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed in the drive 410 as needed so that a computer program read therefrom can be installed into the storage section 408 as needed.

[0200] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 409 and / or installed from a removable medium 411. When the computer program is executed by the central processing unit (CPU) 401, the various functions defined in the present invention are performed.

[0201] It should be noted that specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0202] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings.

[0203] Specifically, the electronic device of this embodiment includes a processor and a memory, the memory is coupled to one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and one or more processors call the computer instructions to enable the electronic device to execute the method provided by the above embodiment.

[0204] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments, or may exist independently and not incorporated into the electronic device. The storage medium carries one or more computer programs, and when executed by a processor of the electronic device, the electronic device implements the methods provided in the above embodiments.

[0205] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features thereof can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention.

[0206] As used in the above embodiments, the term “when…” may be interpreted to mean “if…” or “after…” or “in response to determining…” or “in response to detecting…”, depending on the context. Similarly, the phrases “upon determining…” or “if (stated condition or event) is detected” may be interpreted to mean “if determining…” or “in response to determining…” or “upon detecting (stated condition or event)” or “in response to detecting (stated condition or event)”, depending on the context.

[0207] Those skilled in the art will appreciate that all or part of the process steps in the above-described method embodiments can be implemented by a computer program instructing the relevant hardware. The program can be stored in a computer-readable storage medium. When executed, the program can include the process steps in the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A charge and discharge control method for an elevator energy storage system, characterized in that: include: Acquire real-time operating data of the elevator, including elevator car position, load capacity, traction motor speed, DC bus voltage and current; Determining the instantaneous generated power and power change rate of the elevator traction machine based on the real-time operating data; Determining a predicted power change value of the elevator traction machine within a future preset time period based on the instantaneous generated power and the power change rate; Determine the total amount of generated energy that the elevator energy storage system needs to store within the future preset time period based on the power change prediction value; determine the charging and discharging priority of the elevator energy storage system's supercapacitor, lithium battery, or grid feedback based on the power change rate, the energy storage unit status of the elevator energy storage system, and the grid electricity price period, where the energy storage unit status includes the charge state of the supercapacitor and the health state of the lithium battery; According to the charge and discharge priority, controlling the bidirectional DC / DC circuit of the elevator energy storage system to switch the charge and discharge mode, and adjusting the PWM duty cycle of the braking resistor in the elevator energy storage system, wherein the braking resistor is used to consume the regenerative electric energy generated during the elevator electrical braking process, and the PWM duty cycle is used to control the proportion of the power-on time of the braking resistor to adjust the power consumption of the braking resistor; Based on the charging and discharging mode and the braking resistor PWM duty cycle, the elevator energy storage system is controlled to charge and discharge.

2. The method according to claim 1, characterized in that The step of determining a power change prediction value of the elevator traction machine within a future preset time period based on the instantaneous generated power and the power change rate includes: Generate a multi-dimensional feature data set based on the real-time operation data and the historical power change rate database; Inputting the multi-dimensional feature data set into a pre-trained XGBoost regression model to obtain an initial power change prediction curve including power prediction values ​​at different time points in a future preset time period; Based on the real-time detected traction machine vibration spectrum data, the initial power prediction curve is error compensated through a sliding time window to obtain a power change prediction value of the elevator traction machine.

3. The method according to claim 1, characterized in that The determining, based on the power change prediction value, the total amount of generated energy that needs to be stored by the elevator energy storage system within the future preset time period includes: Resetting the positive values ​​in the power change prediction value to zero and retaining the negative values ​​to obtain a generated power sequence; Performing an integration operation on the generated power sequence within the future preset time period to obtain a theoretically storable amount of electricity; According to the real-time charging and discharging efficiency of the energy storage unit in the elevator energy storage system, the loss compensation is performed on the theoretically storable amount of electricity to obtain the actual total amount of generated energy that needs to be stored.

4. The method according to claim 1, wherein The determining of the charging and discharging priority of the supercapacitor, lithium battery, or grid feedback of the elevator energy storage system based on the power change rate, the energy storage unit status of the elevator energy storage system, and the grid electricity price period includes: determining a fluctuation state of the current load of the elevator traction machine based on the power change rate, wherein the fluctuation state includes a high-frequency transient fluctuation or a steady-state fluctuation; Determining the availability of the energy storage unit based on the state of charge of the supercapacitor and the health status of the lithium battery; Based on the real-time clock and the electricity price database, determine the current electricity price period, which includes a peak electricity period, a valley electricity period, or a normal period; The fluctuation state, the availability of the energy storage unit and the current electricity price period are input into a preset rule library to obtain the charging and discharging priority.

5. The method according to any one of claims 1 to 4, characterized in that The method of controlling the bidirectional DC / DC circuit of the elevator energy storage system to switch the charge and discharge mode according to the charge and discharge priority, and adjusting the brake resistor PWM duty cycle of the elevator energy storage system, includes: Generate a switching signal for a charge / discharge mode according to the charge / discharge priority, wherein the charge / discharge mode includes a supercapacitor priority charging mode, a lithium battery priority discharging mode, or a grid feedback mode; Based on the switching signal, controlling the bidirectional DC / DC circuit to switch the charge and discharge mode; After switching the charge and discharge mode, the brake resistor PWM duty cycle is determined according to the detected real-time deviation of the bus voltage.

6. The method according to claim 5, characterized in that Determining the brake resistor PWM duty cycle according to the detected real-time bus voltage deviation includes: Constructing a sliding mode surface in a sliding mode control algorithm based on the real-time deviation and change rate of the bus voltage; When the absolute value of the sliding mode surface is greater than the sliding mode threshold, performing a sliding mode control calculation to obtain a first duty cycle increment; According to the real-time temperature of the braking resistor, the first duty cycle increment is gradient constrained to obtain a second duty cycle increment; based on the second duty cycle increment, the real-time temperature of the braking resistor and the historical operation frequency, the braking resistor PWM duty cycle is generated.

7. The method according to claim 6, characterized in that The functional expression of the sliding surface includes: Among them, S represents the value of the sliding surface, ΔV represents the real-time deviation of the bus voltage, represents the fractional differential term of ΔV, α is the order of the fractional differential operator, t is time, λ(t) represents the time-varying coupling coefficient, λ0 is the initial coupling coefficient, sign(ΔV) represents the sign function based on ΔV, γ is the potential field intensity coefficient, β is the potential field attenuation coefficient, τ represents the integral time variable, k is the cumulative gain coefficient, and e is a natural constant.

8. An electronic device, characterized in that: including one or more processors and memory; The memory is coupled to the one or more processors, and is configured to store computer program codes, where the computer program codes include computer instructions. The one or more processors call the computer instructions to enable the electronic device to execute the method according to any one of claims 1 to 7.

9. A computer-readable storage medium storing computer instructions, characterized in that: When the computer instructions are executed on an electronic device, the electronic device is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is run on an electronic device, the electronic device is enabled to perform the method according to any one of claims 1 to 7.

Images