Motor Soft-Start Technology: Intelligent Control, Wide Bandwidth Applications, and Energy Efficiency Optimization

Abstract

Direct-starting of industrial motors has problems such as large current impact (five to eight times the rated current), mechanical stress damage, and low energy efficiency. This paper explores the technological innovations in motor soft-start driven by intelligent control and wide-bandgap semiconductors, and constructs a highly reliable and low energy consumption solution. Firstly, based on a material–device–algorithm system framework, a comparative study is conducted on the performance breakthroughs of SiC/GaN in replacing silicon-based devices. Secondly, an intelligent control model is established and a highly reliable system architecture is developed. A comprehensive review of recent literature indicates that SiC devices can reduce switching losses by up to 80%, and intelligent algorithms significantly improve control accuracy. System-level solutions reported in the industry demonstrate the capability to limit current to 1.5–3 times the rated current and achieve substantial carbon emission reductions. These technologies provide key technical support for the intelligent upgrading of industrial motor systems and the dual-carbon goal. In the future, development will continue to evolve in the direction of device miniaturization and other directions.

1. Introduction

Motor soft-start technology intelligently controls the motor startup process, addressing issues caused by traditional direct-starting such as current surges (reaching 5–8 times the rated current I_rated) and mechanical stress. This protects equipment and extends its service life [1]. Its core principles are categorized into voltage-controlled (step-up) and current-controlled (current-limited starting) approaches, achieving smooth acceleration through preset acceleration curves [2]. With advancements in power electronics and AI algorithms, this technology has evolved from basic buck starting to intelligent, integrated systems. It now supports high-voltage, high-power scenarios (e.g., SiC device applications) and complex load adaptation (e.g., fuzzy control) [3,4]. Future trends focus on wide-bandgap semiconductors, adaptive algorithms, and energy efficiency optimization, delivering highly efficient and reliable technical solutions for industrial motor systems. This paper systematically analyzes the principles, key technologies, latest advancements, and development directions, providing a comprehensive reference for engineers.

Modern soft-starters typically employ power electronic devices such as Silicon-Controlled Rectifiers (SCRs) and Insulated-Gate Bipolar Transistors (IGBTs) as switching elements. By adjusting the conduction angle or duty cycle of these devices, the output voltage or current is controlled [5,6]. Figure 1 and Figure 2 show actual photographs of thyristors and IGBTs at different power ratings. Low-voltage systems employ a single thyristor (as shown in Figure 3a), while high-voltage systems require multiple thyristors connected in series (as shown in Figure 3b). The three-phase power supply connects to the motor via a main circuit containing thyristors. Voltage and current detection feedback signals are sent to the control system, which uses a drive circuit to control the thyristor trigger delay angle, gradually increasing the voltage to achieve smooth motor acceleration. Upon reaching the rated speed ω_rated, the motor can be directly connected to the grid via a bypass contactor. The system also supports parameter configuration and status feedback [7].

In thyristor-based voltage control and phase regulation, the effective value of the thyristor output voltage U_out is calculated as follows:

$$U_{\mathrm{out}}=U_{\mathrm{in}}\sqrt{{\displaystyle \frac{1}{2\pi }}\sin (2\alpha )+{\displaystyle \frac{\pi -\alpha }{\pi }}}$$

(1)

where U_in denotes the effective value of the input line voltage, and α represents the thyristor trigger delay angle (rad). Analysis of expression (1) reveals that during voltage ramp startup, adjusting α from π → 0 enables a smooth rise in U_out from 0 to U_in [8].

In the thermal design and component selection of motor soft-starters, it is necessary to estimate the thyristor junction temperature T_j [9], which is defined as follows:

$$T_j=T_a+(R_{th(j-c)}+R_{th(c-a)})\cdot (U_T\cdot I_{avg}+r_T\cdot {\displaystyle I_{rms}^2})$$

(2)

where T_a represents the ambient temperature; Rth_(j−c) denotes the junction-to-case thermal resistance; Rth_(c−a) indicates the case-to-ground thermal resistance; U_T is the thyristor conduction voltage drop; r_T is the on-state resistance; I_avg is the average current; and I_rms is the effective current. The core principle for thyristor design in soft-start technology is rationally matching thermal parameters with the device’s electrical characteristics to ensure its junction temperature T_j does not exceed the standard silicon device tolerance limit of 125 °C. This safeguards long-term operational stability and reliability while mitigating risks of performance degradation or failure caused by overheating.

Motor startup time t_start denotes the total time required for the motor to accelerate from rest to rated speed and is defined as follows:

$$t_{start}={\displaystyle \frac{J\cdot \Delta \omega }{K_T\cdot (U_{ratio}-U_{load})}}$$

(3)

where J represents the load rotational inertia; Δω denotes the target angular velocity difference; K_T is the motor torque constant; U_ratio is the soft-start voltage ratio (i.e., the ratio of output voltage to rated voltage, U_out/U_rated); and U_load is the load torque converted to voltage. From an engineering perspective, increasing the soft-start voltage ratio U_ratio can effectively shorten the startup time t_start. However, this adjustment is accompanied by an increase in current surge. Therefore, practical applications require a reasonable balance between startup efficiency and current stability [10,11].

In current and torque control, the starting transient peak current I_peak must be limited as follows:

$$I_{\mathrm{peak}}={\displaystyle \frac{U_{\mathrm{out}}}{\sqrt{{(R_s+R_r)}^2+{(X_s+X_r)}^2}}}$$

(4)

where R_s represents the stator resistance; R_r represents the rotor resistance; X_s represents the stator leakage reactance; and X_r represents the rotor leakage reactance. The starting transient peak current I_peak quantifies the intensity of current surges during the starting process. Its core control logic involves precisely regulating the output voltage U_out to ensure that the starting transient peak current I_peak does not exceed three times the rated current I_rated (i.e., I_peak ≤ 3I_rated). This approach safeguards starting performance while preventing excessive currents from damaging the motor and power grid [12,13,14].

Soft-start technology typically limits the starting current to a range of 1.5 to 3 times the rated current I_rated, depending on the specific load inertia and the selected current-limiting curve [12,13,14]. For example, fan loads can be started at lower current multiples, while heavy-duty crushers may require the upper limit of this range. Its core principle involves “trading energy time for space”—distributing concentrated electrical energy across the starting cycle to reduce power peaks, thereby mitigating impacts on both the power grid and mechanical systems [14]. Furthermore, the electromagnetic torque T_e is defined as follows:

$$T_e={\displaystyle \frac{3p}{{\omega }_s}}\cdot {\displaystyle \frac{{\displaystyle U_{out}^2}{R}_r/s}{{(R_s+R_r/s)}^2+{(X_s+X_r)}^2}}$$

(5)

where p denotes the number of pole pairs in the motor; ω_s represents the synchronous angular velocity; and s indicates the slip rate (at the initial startup moment, slip rate s = 1). When employing a constant-current startup strategy in soft-start technology, the core optimization objective is to precisely regulate parameters so that the electromagnetic torque T_e exhibits linear growth. This prevents mechanical impact caused by torque transients, enabling smooth motor startup and extending equipment lifespan [15].

Effective thermal management requires controlling energy consumption. The total energy consumption during the startup process is E_start, which is calculated as follows:

$$E_{start}={\displaystyle {\displaystyle {\int }_0^{t_{start}}}3}{\displaystyle I_{rms}^2}(t)R_{s}dt+{\displaystyle {\displaystyle {\int }_0^{t_{start}}}{T}_e}(t)\omega (t)dt$$

(6)

where I_rms (t) represents the stator current effective value function (A); T_e (t) represents the instantaneous electromagnetic torque; and ω (t) denotes the rotor angular velocity. Compared with direct-starting, soft-start technology reduces energy consumption by 15–30%. This energy-saving advantage stems from its effective reduction in the peak stator current RMS value I_rms (t), thereby decreasing reactive power losses and thermal losses during the starting process. This approach enhances energy utilization efficiency while ensuring reliable starting performance [16].

The performance of motor soft-start systems hinges on the integrated application of multiple key technologies. These technologies form the core competitive edge of modern soft-start devices, and a thorough analysis of them is crucial for the design, selection, and application of such equipment.

Power electronic devices and topologies form the foundation of soft-starters. Thyristors have long been the preferred choice for high-voltage soft-starters due to their high-voltage and high-current characteristics. Meanwhile, fully controllable devices such as IGBTs, with their high switching frequencies and flexible control capabilities, are increasingly being adopted in low-voltage or small-to-medium power applications.

Table 1. Comparison of major soft-start technology types.

Technology Type	Representative Device	Control Characteristics	Suitable for
Thyristor soft-starter [17]	Thyristor	Phase control, adjusting the conduction angle	Medium- and high-voltage ac motors
Magnetically controlled soft-starter [18]	Saturation reactor	Magnetic saturation regulation, closed-loop current	High-voltage, high-power motor
Electronic soft-starter [19]	IGBT	Pwm control, multiple modes	Low-voltage precision control
Brushless DC control [20]	MOSFET	Communication configuration, multi-mode	Brushless DC motor

compares the main types of soft-starter technologies.

In terms of topology, beyond the traditional series soft-start circuit topology shown in Figure 4, integrated bypass design has emerged as a new trend. By incorporating the bypass switch directly into the soft-starter unit, these designs achieve higher-power density and simplify installation [21].

The loss ratio η_loss of wide-bandgap semiconductor power modules is expressed as follows:

$${\eta }_{\mathrm{loss}}={\displaystyle \frac{P_{cond}+P_{sw}}{P_{out}}}={\displaystyle \frac{R_{ds(on)}\cdot {\displaystyle I_{rms}^2}+(E_{on}+E_{off})\cdot f_{sw}}{P_{out}}}$$

(7)

where R_{ds (on)} represents the on-resistance (the on-resistance of SiC devices (15 mΩ) is significantly lower than that of Si devices (60 mΩ)); I rms is the effective value of the operating current; E_on and E_off represent the switching energy (the switching energy of SiC devices (0.5 mJ) is only one-quarter that of Si devices (2 mJ)); and f_sw denotes the switching frequency. The breakdown voltage U_br of SiC devices is calculated as follows:

$$U_{br}={\displaystyle \frac{{\displaystyle E_c^2}\cdot {\epsilon }_s\cdot {\epsilon }_0}{2q\cdot N_d}}$$

(8)

where E_c represents the critical breakdown field strength (the critical breakdown field strength of SiC (3 MV/cm) is 10 times that of Si-based materials (0.3 MV/cm)); ε_s is the material dielectric constant (SiC: 9.7, Si: 11.9); N_d is the doping concentration; ε₀ is the vacuum dielectric constant; and q is the elementary charge (fundamental charge). In engineering applications, SiC devices achieve the same breakdown voltage with a thickness only one-tenth that of silicon-based devices.

Advanced control algorithms are central to enhancing soft-start performance. Traditional PID control (as shown in Figure 5a) struggles to meet starting demands under complex load conditions. Modern soft-starters increasingly adopt intelligent control algorithms. For loads with square torque characteristics, such as fans and pumps, the literature proposes using fuzzy control algorithms to adaptively adjust the slope of the limiting current (as shown in Figure 5b). This enables the electromagnetic torque curve to fit the load torque curve, achieving near-constant acceleration starting [15,21].

Zhejiang Xinmai Technology’s three-phase brushless DC motor soft-starter offers at least three configurable soft-start modes, supporting flexible parameter settings for start voltage, frequency, and duration. This addresses the limitations of existing start methods, such as single-mode operation and poor parameter adaptability. High-voltage dry-type magnetically controlled soft-start devices (also known as saturable reactor soft-starters) regulate AC motor current by controlling the magnetic saturation of the reactor core. This technology should not be confused with “magnetron” devices used in microwave applications.

The pump load startup time t_pump is expressed as follows:

$$t_{\mathrm{pump}}={\displaystyle \frac{J}{K_p}}\ln \left({\displaystyle \frac{{\omega }_{\mathrm{rated}}}{{\omega }_{\mathrm{rated}}-0.95{\omega }_{nl}}}\right)$$

(9)

where K_p is the pump load coefficient; ω_nl is the no-load rotational speed; and ω_rated is the rated rotational speed. Since the torque (T_L) of the pump load is proportional to the square of the rotor angular velocity (ω) (T_L∝ω²), this nonlinear load characteristic necessitates a composite control strategy of “voltage ramp + current limiting” during the starting process. This approach ensures smooth acceleration, avoids excessive current surges, and guarantees stable startup and efficient operation of pump equipment [22].

The optimal starting current value I_{start_opt} for the fan is calculated as follows:

$$I_{\mathrm{start}\_\mathrm{opt}}=I_{\mathrm{rated}}\sqrt[3]{{\displaystyle \frac{T_{\mathrm{L}\_\mathrm{rated}}}{T_{\mathrm{e}\_\max }}}}$$

(10)

where T_{L_rated} denotes the rated load torque and T_{e_max} represents the motor’s maximum torque. The key control strategy of soft-start technology is to precisely set the starting current (I_start) to its optimal value (I_{start_opt}). This ensures that the starting torque requirement is met while minimizing mechanical stress on the motor and transmission system, thereby achieving smooth startup and extending equipment lifespan. The current reduction rate η_I is commonly used to evaluate current suppression effectiveness and is defined as follows:

$${\eta }_I=\left(1-{\displaystyle \frac{I_{\mathrm{soft}\_\mathrm{start}}}{I_{\mathrm{direct}}}}\right)\times 100\%$$

(11)

where I_direct represents the peak current during direct-starting and I_{softs_sart} denotes the peak current during soft-starting. For high-voltage solid-state soft-starting technology, its typical current reduction efficiency can exceed 60% (i.e., η_I > 60%). Therefore, soft-starting technology can significantly reduce current peaks during the starting process, effectively mitigating impacts on the power grid and motor equipment while enhancing the safety and stability of the starting process [23]. The average torque fluctuation coefficient K_T serves as another metric for evaluating soft-starts and is defined as follows

$$K_T={\displaystyle \frac{\max (T_e)-\min (T_e)}{T_{\mathrm{avg}}}}\times 100\%$$

(12)

where T_avg represents the average torque during the starting process and max (T_e) and min (T_e) denote the maximum and minimum values of the electromagnetic torque T_e. The core optimization objective is that, in precision control applications, the average torque fluctuation coefficient K_T must be controlled within 10% (i.e., K_T < 10%). This ensures stable torque output during the starting process, meeting the stringent requirements for smooth operation in high-precision equipment and reducing mechanical vibration or control errors caused by torque fluctuations [24].

System protection and reliability technologies are crucial for the long-term stable operation of soft-starters [25]. Closed-loop fault detection circuits are increasingly integrated into the triggering stage to monitor thyristor conduction states in real time, thereby preventing cascading failures [25]. Jiuquan Iron and Steel Group’s Hongxing Steel implemented control circuits that disconnect control voltage after soft-starter-assisted startup, resolving issues of equipment malfunction shutdowns and long-term component aging due to continuous energization. Shenzhen Yanzhi Electric’s patented soft-start circuit leverages coordinated first-current detection, comparison, and reference source circuits to overcome the challenges of impact and damage caused by direct motor starting. Typically, the time deviation Δt between actual and theoretical triggering is utilized to diagnose thyristor triggering anomalies, where Δt is defined as follows:

$$\Delta t={\displaystyle \frac{1}{2\pi f}}\left|\arcsin \left({\displaystyle \frac{U_{\mathrm{th}}}{U_{\mathrm{p}}}}\right)-{\alpha }_{\mathrm{ref}}\right|>{\tau }_{\mathrm{safe}}$$

(13)

where f represents the grid frequency; Uth denotes the thyristor threshold voltage; U_p signifies the peak phase voltage of the grid; αref indicates the theoretical trigger delay angle; and τ_safe represents the safety tolerance time, typically valued at 10 μs. The core protection logic operates as follows: when the time deviation (Δt) between the actual triggering and theoretical triggering exceeds the safety tolerance time (τ_safe), the system identifies a triggering anomaly and activates the redundant thyristor or switches to the standby phase. This prevents current surges or equipment failures caused by triggering anomalies, ensuring the reliability and safety of the soft-start process.

Heat dissipation and structural design technologies directly impact the power density and reliability of soft-starters [26]. Shanghai Renuoer Technology’s intelligent control devices feature dedicated heat sinks to ensure stable thyristor operation, while high-voltage solid-state soft-starter cabinets achieve compact layouts through optimized structural design to reduce user maintenance costs; Xiangyang Yuanchuang Electric’s YCK series electromagnetic soft-starters are suitable for soft-starting medium-to-large AC motors across voltage levels of 380 V, 6 kV, and 10 kV, and power ratings from 100 kW to 20 MW. Their contactless, wear-free characteristics deliver superior reliability in harsh environments.

Communication and intelligent technologies are becoming standard features in modern soft-starters. High-voltage dry-type magnetic control soft-starters equipped with advanced microelectronic control systems enable remote monitoring and communication functions. Currently, the vast majority of soft-starters incorporate communication interface modules that can communicate with external controllers to configure motor operating parameters. This trend toward intelligence transforms them from isolated electrical equipment into smart nodes within the industrial Internet of Things, laying the foundation for predictive maintenance and energy efficiency optimization [27].

While the boundaries between soft-starters and Variable Frequency Drives (VFDs) are increasingly blurring due to the advancement of power electronics, distinct application ecosystems remain. It is therefore essential to define the scope of this review to differentiate between these technologies.

Applicable Motor Types and Ranges: This review primarily focuses on soft-start technologies for asynchronous (induction) motors, which remain the workhorse of industry due to their varied dynamic regimes and robustness, as analyzed in recent comprehensive studies by Konuhova et al. [3]. However, significant attention is also given to synchronous motors, where fault detection and startup synchronization are critical [4], as well as emerging applications in brushless DC (BLDC) motors [20]. The scope covers a wide power spectrum, from low-voltage commercial units (380 V, <100 kW) to high-voltage industrial drives (10 kV, >20 MW).

Soft-Starter vs. VFD Positioning: A soft-starter is the preferred technological choice over a VFD in the following applications: Speed control is not required: the process runs at a fixed rated speed (e.g., ventilation fans and steady-state pumps). Cost and size are constrained: soft-starters are typically 30–50% of the cost and volume of a comparable VFD.

Harmonic mitigation is a priority: unlike VFDs, which generate significant harmonics during the entire operation, thyristor-based soft-starters are bypassed after startup, resulting in near-zero harmonic distortion during the run state [28].

To provide a clear selection guide for engineers,

Table 2. Comparative analysis of motor starting technologies.

Feature	Direct-on-Line (DOL)	Star–Delta (Y-Δ)	Autotransformer	Soft-Starter	Variable Frequency Drive (VFD)
Starting Current	Very High (5–8 Irated)	High (2–3 Irated)	Medium (1.5–3 Irated)	Low/Adjustable (1.5–3 Irated)	Very Low (<1.5 Irated)
Torque Control	Uncontrolled (High Shock)	Step Change (Mechanical Shock)	Fixed Steps	Smooth/Linear Ramp	Precise Full Range Control
Speed Regulation	None	None	None	None (Startup/Stop Only)	Full Continuous Control
Harmonics	Negligible	Negligible	Negligible	Low (Startup Only)	High (Requires Filters)
System Cost	Low	Low	Medium	Medium	High
Physical Footprint	Small	Medium (2 contactors)	Large (Bulky Transformer)	Compact	Large (Heatsinks Required)
Energy Efficiency	High	High	Medium (Heat Loss)	High (Bypass Mode)	Medium (Switching Losses)
Typical Application	Small motors (<7.5 kW)	Cost-sensitive HVAC	Old MV Grids	Pumps, Conveyors, and Crushers	Precision Process Control

presents a comparative analysis of common starting methods.

2. Recent Advances in Motor Soft-Start Technology

In recent years, with the deepening advancement of Industry 4.0, smart manufacturing, and green energy concepts, motor soft-start technology has achieved significant breakthroughs in intelligence, integration, and efficiency. These innovations have not only enhanced the performance of the devices themselves but also expanded their application scenarios, providing superior solutions for industrial energy conservation and equipment protection.

2.1. Intelligent and Networked Control

Intelligent and networked control represent the primary development direction for soft-start technology. Modern magnetic-controlled soft-starters are often equipped with advanced microelectronic control systems to enable remote monitoring and communication, allowing users to operate them in real time via smart devices or network platforms [28]. Zhejiang Xinmai Technology’s patented soft-starter integrates a communication interface module, supporting configuration and storage of motor parameters through external controllers to enable customization of different starting modes. Shanghai Renol Technology’s intelligent control device employs current transformers for real-time current monitoring to ensure motor operational safety. Its compact design accommodates space-constrained environments, advancing soft-starters from basic motor control equipment to intelligent terminals with data acquisition, analysis, and remote management capabilities [29]. On the international front, leading manufacturers have significantly advanced the digitalization of soft-starters. ABB’s PSTX series, for instance, integrates detachable keypads with fieldbus communication (Modbus/Profibus) and built-in motor protection functions, enabling seamless integration into Distributed Control Systems (DCSs). Schneider Electric’s Altistart 480 features cybersecurity enhancements and native integration with digital twin platforms (EcoStruxure), allowing operators to simulate startup curves and predict thermal behavior before physical deployment. These global solutions emphasize not just the starting process but the complete lifecycle management of the motor asset [30,31,32,33].

The fuzzy control expression for dI_lim/dt in torque fitting for fan-type loads is calculated as follows:

$${\displaystyle \frac{d{I}_{lim}}{dt}}=K_f\cdot {\displaystyle \frac{\mu (T_L,\Delta \omega )}{\max \left(\left|\frac{d{T}_L}{dt}\right|,ϵ\right)}}$$

(14)

where I_lim represents the current limit value; K_f denotes the fuzzy proportional coefficient determined by the expert rule base; T_L characterizes the load torque; Δω indicates the rotational speed deviation; μ (T_L,Δω) is the membership function of the matching degree (with a value range of 0 to 1); and ϵ is the anti-zero constant, typically set at 0.1 N·m/s. The core control objective is to dynamically adjust the current slope using fuzzy control strategies and membership functions. This approach ensures that the difference between electromagnetic torque and load torque approximates the product of rotational inertia and angular acceleration. This achieves constant acceleration motor startup, ensuring smoothness and controllability throughout the starting process [30,31].

The load recognition accuracy A_c based on multi-modal AI strategies is defined as follows [32]:

$$A_{\mathrm{c}}={\displaystyle \frac{1}{n}}{\displaystyle {\displaystyle \sum _{i=1}^n}\mathbb{I}}\left(‖{\displaystyle \frac{J_{\mathrm{pred}}-J_{\mathrm{real}}}{J_{\mathrm{real}}}}‖<0.1\right)$$

(15)

where J_pred denotes the predicted rotational inertia; J_real denotes the actual rotational inertia; n denotes the number of test samples and the number of load types in the validation dataset (covering fans/pumps/crushers, etc.); and II denotes the indicator function, a logical function determining prediction accuracy. It outputs 1 when the relative error is <10%, and 0 otherwise, i.e., expressed as follows:

$$\mathbb{I}(x)=\left\{\begin{array}{l}\begin{array}{cc}1& x<10\%\end{array}\\ \begin{array}{cc}0& x\ge 10\%\end{array}\end{array}\right.$$

(16)

J represents the inertia quantity resisting changes in the rotational speed of the load and is defined as follows

$$T_e-T_L=J{\displaystyle \frac{d\omega }{dt}}$$

(17)

where T_e represents the electromagnetic torque; T_L denotes the load torque; and ω signifies the motor speed. Different load types exhibit varying effects on soft-starting are as follows: high-inertia loads (such as ball mills) require gradual torque ramp-up to prevent mechanical shock, while low-inertia loads (such as fans) can utilize rapid starting methods to shorten the startup time.

Figure 6 illustrates the flowchart of the AI-based prediction mechanism. It first acquires the starting current signal and then converts the time-domain signal to the frequency domain via Fast Fourier Transform (FFT) spectral analysis. Subsequently, key features are extracted, including the fundamental-to-harmonic ratio (H5/H1) reflecting load damping characteristics (e.g., pumps > 0.2, crushers < 0.1) and the fundamental frequency current change rate per unit time (dI₁/dt), which correlates with inertia magnitude (high J loads exhibit low slopes). Subsequently, these features are integrated into a one-dimensional convolutional neural network (1D-CNN). The input layer of this network consists of a 50-dimensional spectrum vector spanning 0–2500 Hz. This vector undergoes processing through three 1D convolutional kernels with a width of 5 and a stride of 1, followed by computational operations in the output layer according to the following formula [33,34].

$$J_{\mathrm{pred}}=W\cdot \sigma \left(Conv\left(S\right)\right)+b$$

(18)

where S represents the spectral vector. The final output is the J_pred prediction result, achieving a complete workflow from current signal acquisition to intelligent prediction. This data-driven approach is employed for motor soft-start condition monitoring and performance evaluation.

Physics-Assisted Control Mechanism: The connection between the AI-extracted features and the physical control law is established through a physics-assisted hybrid modeling approach. Unlike Physics-Informed Neural Networks (PINNs) which typically embed partial differential equations directly into the network’s loss function, the proposed method utilizes physical motion equations to pre-structure feature selection and guide fuzzy inference rules [35].

The error threshold setting for load identification accuracy A_c adopts a hybrid modeling strategy based on both physical principles and industry standards [36]. From a physical perspective, the relationship between the rotational inertia error δJ and the startup time deviation Δt is expressed as follows:

$$\Delta t=t_{\mathrm{start}}\left|\sqrt{{\displaystyle \frac{J_{\mathrm{real}}+\delta J}{J_{\mathrm{real}}}}}-1\right|$$

(19)

where t_start denotes the motor start time, which is a key indicator for evaluating the performance of soft-start systems. It determines the magnitude of current surge intensity and mechanical stress during the startup process. Regarding industry standards, IEC 60034-31 stipulates that motor startup time error should be <7%, while a 10% inertia error provides ample margin for this requirement [37,38,39].

As shown in

Table 3. Overview of recent patents on representative soft-start technologies.

Control Strategy	Complexity	Comparison with Load Variations	Setup	Typical Application Suitability
Traditional PID	Low	Poor (Requires re-tuning if J changes)	High (Manual tuning required)	Fixed loads (e.g., ventilation fans)
Self-Tuning PID	Medium	Moderate (Slow convergence)	Medium (Auto-tuning phase)	Standard pumps/conveyors
I−t Curve ID	Medium	Moderate (Offline/static)	Medium (Needs historical data)	Loads with predictable cycles
Proposed (Physics-Assisted CNN + Fuzzy)	High	Excellent (Real-time J estimation)	Low (Self-adaptive)	Complex/variable Loads

, while traditional methods such as self-tuning PID and parameter identification based on I-t curves are effective for fixed loads, they exhibit significant limitations in complex industrial scenarios.

Limitations of Traditional Methods: Standard PID controllers often struggle with nonlinear loads or time-varying systems, leading to overshoot or mechanical shock [40]. Simple I-t curve identification typically requires offline learning or specific startup conditions, making it unsuitable for real-time variations.

Necessity of the Proposed Route: The proposed “1D-CNN + fuzzy” scheme, although computationally more complex, offers a critical advantage: robustness against unmodeled dynamics. The 1D-CNN can extract inertia features from noisy current signals within milliseconds, allowing the fuzzy controller to preemptively adjust the starting curve before the system saturates. This trade-off—accepting higher algorithm complexity to achieve “commissioning-free” adaptability and reduced mechanical stress—is justified in high-value equipment where startup failure costs far exceed the computational cost.

Figure 7 illustrates the effect of design choices on t_start [22,35].

Figure 7 illustrates three starting modes and their corresponding durations, which must comply with industry standards: The voltage ramp mode initiates startup through gradual voltage increase, taking the longest duration of approximately 12 s. It is suitable for scenarios requiring smooth startups to avoid current surges. The constant-current control mode maintains a steady starting current, with a moderate duration of about 9 s, striking a balance between starting current and speed. The torque control mode directly regulates output torque, achieving the fastest startup in approximately 7 s and is suitable for scenarios requiring rapid startup and high torque response from the equipment. Regarding industry standards, IEC 60034-12 stipulates that the startup process must not cause motor temperature rise exceeding insulation class limits. For instance, when I_start = 3I_rated and t_start ≤ 67 s, the constraint expression for the maximum startup current multiple multiplied by time must satisfy the following:

$$I_{\mathrm{start}}\times t_{\mathrm{start}}\le I_{\mathrm{rated}}\times 200\mathrm{s}$$

(20)

GB/T 12668.3 specifies specific load requirements: for pump loads, t_start > 10 s to prevent water hammer effects; for crusher loads, t_start > 15 s to reduce mechanical impact.

2.2. High-Voltage, High-Power Soft-Start Technology

High-voltage, high-power soft-start technology continues to innovate to meet the demands of heavy industries such as power generation and chemical processing. Among these, high-voltage dry-type magnetic-controlled soft-start devices regulate AC motor current through saturated reactors, offering both high performance and low cost. They provide optimized current waveforms to reduce grid harmonic pollution while demonstrating robust tolerance to sudden overloads [36,37]. The saturated reactor regulates starting current I_motor by adjusting magnetic saturation, where it can be calculated as follows

$$I_{\mathrm{motor}}=k\cdot {\displaystyle \frac{U_{\mathrm{grid}}}{X_{\mathrm{sat}}(\Phi )}}$$

(21)

where k is the device structure coefficient (dimensionless, related to reactor design); U_grid is the grid input voltage; and X_sat (Φ) is the dynamic reactance of the saturated reactor, whose value varies with magnetic flux Φ. The higher the magnetic saturation, the smaller X_sat becomes, and the greater the current.

The high-voltage solid-state soft-starter cabinet precisely controls the starting current I_start, reducing it to a range of 1.5 to 3 times the rated current I_rated, which is adjustable:

$$1.5I_{\mathrm{rated}}\le I_{\mathrm{start}}\le 3I_{\mathrm{rated}}$$

(22)

Soft-start technology prevents excessive inrush currents during motor startup and reduces peak grid loads. Chongqing Zongshen General Power Machinery’s patented high-current soft-start circuit employs a segmented control strategy to ensure smooth startup during sudden high-current loading, mitigating potential damage from excessive power surges during maximum-power startup.

Similarly, in the field of heavy-duty applications, Rockwell Automation (Allen-Bradley) has developed the SMC-50 controller with advanced linear speed acceleration and torque control, which is widely used in global mining and petrochemical industries to prevent conveyor belt slippage. Siemens’ SIRIUS 3RW55 series introduces hybrid switching technology that combines the durability of mechanical contacts with the precision of power electronics, effectively handling high-inertia loads while minimizing heat dissipation [38,41].

Figure 8 shows actual images of mining-grade soft-start thyristor modules (FXR7-ST-KP800A4200V, FXR8-ST-KP1250A4200V) manufactured by Shanghai Nantai Rectifier Co., Ltd.

Figure 9 shows the actual valve assembly of the SWS-H series high-voltage motor solid-state soft-starter. It covers voltage ratings of 3.3 kV, 6.6 kV, and 10 kV, with power ratings ranging from 200 kW to 30 MW. Considering the high switching speed of SiC devices, the parasitic inductance and resistance are sensitive to skin and proximity effects. To ensure simulation accuracy, the solution frequency in Ansys Q3D was explicitly set to 30 MHz. This frequency selection is based on the equivalent frequency of the switching transient. For the SiC MOSFETs used in this study, the typical rise time is approximately 10 ns, corresponding to a spectral content significantly higher than the fundamental switching frequency. Extracting parameters at this frequency ensures that the simulation captures the high-frequency AC impedance characteristics that dominate voltage overshoot and ringing events rather than the DC or low-frequency values.

2.3. Customization and Scenario Adaptation Technology

As shown in

Table 4. Comparison of Technical Features and Application Advantages of Peer Motor Soft-start Devices.

Examples of Peer Devices	Technical Features	Application Advantages
Three-phase AC voltage self-bypass motor soft-start control unit	Integrated bypass switch with real-time current detection [37]	Compact structure and easy to assemble and maintain
Three-phase brushless DC motor soft-starter	Three configurable soft-start modes	Highly adaptable, stable, and quiet
Soft-starter with controllable silicon triggering abnormal protection	Fault detection feedback closed-loop control	Enhance system stability and security
High-current soft-start circuit	Segmented current control [38]	Handling sudden load conditions
Asynchronous motor soft-start control method and system	Fuzzy control algorithm [39]	Suitable for fan and pump loads

, specialized and scenario-adaptive technologies have developed optimized solutions for diverse application scenarios. For loads with square torque characteristics, such as fans and pumps, the load torque T_L is expressed as follows:

$$T_L=K_T\cdot {\omega }^2+T_0$$

(23)

where K_T denotes the torque proportional coefficient; ω represents the motor speed; and T₀ indicates the static friction torque. For fan/pump loads, torque is proportional to the square of the speed. During startup, T_L is extremely low, while at the rated point it rises to its maximum value.

The fuzzy control algorithm adaptively modifies the slope of the limiting current as follows:

$${\displaystyle \frac{d{I}_{\lim }}{dt}}=K_c\cdot \mu (\Delta T)\cdot \mathrm{sgn}\left({\displaystyle \frac{d\omega }{dt}}\right)$$

(24)

where I_lim represents the soft-start current limit value; K_c denotes the control gain coefficient; μ (ΔT) is the membership function (with a value range of 0 to 1), where ΔT = T_e − T_L; and sgn (.) indicates the sign function, taking +1 during acceleration and −1 during deceleration. By dynamically adjusting the current rise rate through the membership function μ (ΔT), the electromagnetic torque T_e tracks the load torque T_L. This enables the electromagnetic torque curve to fit the load torque curve, achieving near-constant acceleration starting. It significantly reduces current surges, torque surges, and energy losses during motor startup. Zhejiang Xinmai Technology’s brushless DC motor soft-starter offers multiple configurable starting modes, addressing the limitations of existing single-mode approaches with poor parameter adaptability. It enables stable, low-noise motor starts suitable for diverse applications, including home appliances and industrial equipment. Jiuquan Iron and Steel Group Hongxing Steel’s related patent improves control circuits to ensure that the soft-starter enters a cold, de-energized state after completion, specifically resolving issues of erroneous shutdowns during operation and component aging caused by prolonged energization.

To ensure the engineering feasibility of power module layouts, strict geometric constraints, including electrical clearance and component non-overlap rules, are incorporated into the optimization process. We adopt a hybrid constraint-handling strategy within the NSGA-II framework. Specifically, validity checking is performed during the population initialization phase to strictly avoid generating individuals with component overlaps or boundary violations. For constraints that are computationally expensive to verify or rectify instantly, a static penalty function approach is employed during the evolutionary process. Individuals violating these design rules are assigned a large penalty value to their objective functions, effectively lowering their rank during the non-dominated sorting. This mechanism ensures that the population naturally evolves towards the feasible region while adhering to all design rules.

2.4. Energy-Saving and Environmental Protection Technologies

Energy-saving and environmental protection technologies are highly valued in the soft-starting field. High-voltage solid-state soft-start cabinets effectively reduce motor energy consumption and lower carbon emissions by precisely controlling starting currents, minimizing starting stresses, and achieving energy savings under light loads [40]. The energy consumption ratio η_E is defined as the ratio of soft-start energy consumption E_soft to direct-start energy consumption E_direct, expressed as follows:

$$\begin{array}{l}{\eta }_E={\displaystyle \frac{E_{\mathrm{soft}}}{E_{\mathrm{direct}}}}={\displaystyle \frac{{\displaystyle {\displaystyle {\int }_0^{t_s}}{\displaystyle I_{\mathrm{soft}}^2}}(t)R_{s}dt+P_{\mathrm{core}}}{{\displaystyle {\displaystyle {\int }_0^{t_d}}{\displaystyle I_{\mathrm{direct}}^2}}(t)R_{s}dt+P_{\mathrm{core}}}}\\ \approx 0.65~0.75\end{array}$$

(25)

where E_soft and E_direct represent the total energy consumption for soft-start and direct-start, respectively; I_soft (t) denotes the time-varying soft-start current function; I_direct indicates the direct-start current (typically 5–8 times I_rated); Rs is the stator resistance; P_core is the core loss, approximated as constant; and t_s and td denote the soft-start and direct-start times, respectively.

Magnetic-controlled soft-starters effectively reduce energy consumption and enhance energy utilization efficiency by employing advanced energy-saving technologies such as power factor correction and reactive power compensation [11]. The reactive power Q_c requiring compensation for magnetic-controlled soft-starters is calculated as follows:

$$Q_c=P_m\left(\sqrt{{\displaystyle \frac{1}{P{\displaystyle F_1^2}}}-1}-\sqrt{{\displaystyle \frac{1}{P{\displaystyle F_2^2}}}-1}\right)$$

(26)

where P_m denotes the motor’s active power; PF₁ represents the power factor before compensation (typically ranging from 0.4 to 0.6); and PF₂ indicates the target power factor (>0.95). After compensation, the line loss is reduced by ΔP_L, which is expressed as follows

$$\Delta P_L={\displaystyle \frac{{(Q_c)}^2{R}_{line}}{U^2}}$$

(27)

where R_line represents the line resistance and U represents the line voltage. The expression for the power factor improvement ΔPF is calculated as follows:

$$\Delta PF={\displaystyle \frac{Q_c/S}{\sqrt{1+{(Q_c/S)}^2}}}$$

(28)

where S represents the apparent power. The carbon emission reduction per start, ΔCO₂, is calculated as follows:

$$\Delta C{O}_2=(E_{\mathrm{direct}}-E_{\mathrm{soft}})\times C_{\mathrm{grid}}\times {10}^{-3}$$

(29)

where E_direct and E_soft represent direct/soft-start energy consumption, respectively, and C_grid denotes the grid carbon emission factor. The total annual carbon reduction T_CO2 is expressed as follows:

$$T_{C{O}_2}=\Delta C{O}_2\times N_{\mathrm{start}}\times 365$$

(30)

where N_start represents the average number of daily starts. To quantify the environmental benefits, a typical industrial scenario is considered: a 1000 kW high-voltage motor with an average of five starts per day. Assuming a grid carbon emission factor of 0.58 kg CO₂/kWh and comparing direct-starting with soft-starting, the energy savings per start is calculated using Equation (25). The theoretical calculation based on Equation (30) indicates that, for such heavy-duty industrial applications, the annual carbon emission reduction can reach approximately 15 t.

Shenzhen Yanzhi Electric Technology Co., Ltd.’s patent titled “A Soft-Start Circuit and Motor” effectively addresses energy waste and equipment damage caused by direct motor starts through coordinated operation between the circuit and reference source circuit. This innovation also contributes to achieving green manufacturing and low-carbon production.

2.5. Reliability Enhancement Technology

Reliability enhancement technologies have achieved breakthroughs in multiple aspects of soft-starters. Chint Electric’s thyristor trigger anomaly protection technology establishes a closed-loop control system by integrating fault detection circuits within soft-starters, significantly boosting system stability and safety. Hongxing Iron & Steel of Jiuquan Iron & Steel Group’s cold-state operation control circuit addresses component aging caused by prolonged energization in traditional soft-starters, extending equipment lifespan. High-voltage solid-state soft-starter cabinets incorporate comprehensive and reliable protection functions, such as overload and short-circuit protection. These features activate promptly when motor or related equipment malfunctions, preventing equipment damage and energy waste. Collectively, these technological innovations enhance the overall reliability of soft-starter devices while reducing user maintenance costs.

The system total failure rate (λ_sys) is defined as follows:

$${\lambda }_{\mathrm{sys}}={\displaystyle \sum _{i=1}^n}{\lambda }_i\cdot {\pi }_T\cdot {\pi }_E+{\lambda }_{\mathrm{sw}}\cdot e^{{\displaystyle \frac{E_a}{k}}\left({\displaystyle \frac{1}{T_j}}-{\displaystyle \frac{1}{T_0}}\right)}+{\lambda }_{\mathrm{ctrl}}\cdot {\pi }_p$$

(31)

where λ_sys is composed of the failure rates of passive components, power devices, and control circuits in sequence. λ_i represents the base failure rate of the i th component; π_T denotes the temperature acceleration factor (where π_T = e^0.1(Tj−85), with T_j being the junction temperature); π_E is the environmental stress factor, set to 6.0 for industrial environments; λ_sw indicates the failure rate of switching devices; E_a is the activation energy (0.8 eV for SiC devices, 0.4 eV for Si-based devices); and k is the Boltzmann constant (k = 8.617 × 10⁻⁵ eV/K). The core design principle is that, by rationally matching component parameters and controlling environmental stresses, the total system failure rate (λ_sys) must be ensured. This metric corresponds to the Mean Time Between Failures (MTBFs), guaranteeing long-term high-reliability operation of the system [42].

2.6. Material and Process Innovation

Material and process innovations underpin advancements in soft-start technology [41]. Innovative designs enable high-voltage dry-type magnetic-controlled soft-starters to adapt to motors across different voltage levels. Their compact, maintenance-free structure significantly reduces user operational costs. Shanghai Renuoer Technology’s intelligent control devices optimize heat sink design, ensuring thyristor stability in high-temperature environments. These material and process enhancements have substantially improved modern soft-starters in power density, environmental adaptability, and service life.

3. Application Areas and Benefits of Soft Motor Starting Technology

Relying on significant technical advantages, motor soft-starting technology has been widely used in many fields and has generated considerable economic and social benefits. With continuous technological innovation, its application scenarios are expanding from traditional heavy industry to emerging green energy fields, where it is playing an increasingly important role.

3.1. Areas of Application for Soft Motor Starting Technology

3.1.1. Power and Energy

The power and energy industry is the main application area of high-voltage soft-starting technology. In order to reduce the impact of high-voltage motor starting on the power grid, the motor starting current I_start needs to meet the requirements of grid stability constraints as follows:

$$I_{\mathrm{start}}\le k_{\mathrm{grid}}\cdot I_{\mathrm{grid-rated}}$$

(32)

where k_grid is the grid impact coefficient (dimensionless, set according to the capacity of the plant power system, usually ≤3.5), and I_grid-rated for the rated current of the grid, to ensure that the starting process does not affect the grid voltage stability.

In power plants, the use of soft-starting devices for large fans, pumps, compressors, and other high-voltage motors can effectively reduce the impact of starting current on the power grid and avoid affecting the stability of the plant power system [43]. For sudden overload demands of key equipment in the power industry, the overload capacity of the device needs to satisfy the following:

$$I_{\mathrm{overload}}\le k_{\mathrm{over}}\cdot I_{\mathrm{device-rated}}, \mathrm{and} t_{\mathrm{over}}\le t_{\mathrm{allow}}$$

(33)

where I_overload is the sudden overload current; k_over is the overload coefficient; I_device-rated is the rated current of the device; t_over is the actual overload duration; and t_allow is the allowable overload duration of the device.

High-voltage dry-type magnetically controlled soft-starters (such as the 10 kV/1.7 MW high-voltage dry-type magnetically controlled integrated soft-starter cabinet of Hubei Zhongsheng Electric) are especially suitable for applications involving key equipment in the electric power industry, by virtue of their strong ability to withstand sudden overloads. In the field of power transmission and transformation, soft-start technology can guarantee smooth startup of auxiliary equipment such as the transformer cooling systems and oil pumps, thereby prolonging the service life of the equipment. In the field of renewable energy, yaw and pitch control systems in wind farms are also gradually adopting advanced soft-start solutions to meet the demand for reliable operation in harsh environments [44].

3.1.2. Petrochemical Industry

The petrochemical industry has a rigid demand for soft-starting technology [45]. In order to avoid water hammer effects or sudden pressure changes in piping systems during soft-starting process, it is necessary to control the rate of change in the motor speed Δn/Δt, i.e., can be expressed as follows

$${\displaystyle \frac{\Delta n}{\Delta t}}\le k_p\cdot {\displaystyle \frac{P_{\max }-P_{\mathrm{working}}}{\rho \cdot v\cdot L}}$$

(34)

where k_p is the pipeline safety factor; P_max is the maximum pressure of the pipeline; P_working is the working pressure of the equipment; ρ is the density of the medium; v is the medium flow rate; and L is the length of the pipeline.

The smooth acceleration and soft stop function of the high-voltage solid-state soft-starting cabinet can effectively solve these problems, avoiding the possible problem of wheezing associated with traditional starting equipment. The constraint expression for smooth acceleration current of the high-voltage solid-state soft-starting cabinet is expressed as follows:

$$I_{\mathrm{start}}\le \min (3I_{\mathrm{rated}},I_{\mathrm{pipe-safe}})$$

(35)

where I_start is the motor starting current; I_rated is the motor rated current; and I_pipe-safe is the maximum allowable starting current based on pipeline safety, which is deduced from the pipeline pressure capacity to ensure that there is no sudden change in pressure during acceleration.

In addition, the chemical industry has a lot of explosion-proof requirements. Magnetically controlled soft-starters (such as Zhejiang Nerong Power Equipment’s high-voltage magnetically controlled soft-starter cabinet (NRRQVC), rated voltage of 6 kV/10 kV and rated power of 250 kW–5 MW) are particularly suitable for flammable and explosive occasions by virtue of their non-contact, non-sparking characteristics. Shenzhen Research Intelligence’s patented soft-starter circuits, which accurately control starting current, is are also suitable for chemical production scenarios that require high equipment protection.

3.1.3. Metallurgy and Mining

The harsh working conditions of metallurgical and mining industries put special requirements on soft-starting devices [46]. Crushers, ball mills, and other equipment in mines have large load inertia and are difficult to start, while traditional starting methods are prone to overheating of the motor or damage to the mechanical drive system. The minimum starting torque T_min for large inertia loads is calculated as follows:

$$T_{\min }=J{\displaystyle \frac{d\omega }{dt}}+T_L+T_f$$

(36)

where J is the total rotational inertia of the system, the typical value for ball mills is 500~5000 $\mathrm{kg}\cdot {\mathrm{m}}^2$; dω/dt is the target angular acceleration, usually set to 0.05~0.2 ω_n/s (to avoid mechanical shock); T_L is the load torque, the crusher no-load is about 15%T_rated (rated torque); and T_f is the friction torque, which may reach up to 20%T_rated in mining equipment.

Zongshen Power’s patented high-current soft-starting circuit can ensure the successful realization of soft-starting under sudden high-current loading, which is especially suitable for heavy load starting occasions. The high-temperature and dusty environment of the metallurgical industry requires extremely high environmental adaptability from soft-starters, and magnetically controlled soft-starters can operate stably under harsh conditions, thereby precisely meeting this demand; the cold-running control circuit developed by JISCO Hongxing Iron and Steel Group also solves the long-term reliability problem of soft-starters in the continuous production environment of the metallurgical industry.

3.1.4. Water Treatment and Municipal Engineering

Soft-start technology is widely used in water treatment and municipal engineering fields to optimize equipment operation. Pumps, as core equipment of water treatment systems, benefit from the use of soft-start technology by effectively avoiding water hammer effects and protecting piping systems [47]. For fans, pumps, and other loads with square torque characteristics, the latest asynchronous motor soft-start control methods, with the help of fuzzy control algorithms and adaptive change in the slope of the limiting current, enable the electromagnetic torque curve to match the load torque curve, thereby achieving an approximate constant acceleration starting. Ventilation systems, elevators, and other equipment in the municipal engineering field are also adopting more and more soft-start technology to improve comfort and energy efficiency, among which the compact intelligent control device of Shanghai Reynolds Technology is especially suitable for municipal application scenarios with limited space.

3.1.5. Manufacturing and Machinery

The manufacturing and mechanical equipment field is accelerating the popularization of soft-starting technology. In the manufacturing machinery and equipment field, the deepening development of soft-starting technology has spawned, and the peak impact force F_shock can be expressed as follows:

$$F_{\mathrm{shock}}=J{\displaystyle \frac{d\omega }{dt}}\cdot {\displaystyle \frac{1}{r_{gear}}}+F_0$$

(37)

where J is the load moment of inertia, the typical value for machine tool spindle is 0.5~5 kg × m²; dω/dt is the angular acceleration; r_gear is the gearbox reduction ratio; and F₀ is the preload force. The spindle speed fluctuation rate δ_ω is expressed as follows:

$${\delta }_{\omega }={\displaystyle \frac{\max (\omega )-\min (\omega )}{{\omega }_{\mathrm{ref}}}}\times 100\%$$

(38)

where the rotational speed fluctuation rate δ_ω is based on the set rotational speed ω_ref; ordinary machine tools require δ_ω < 0.5%, while precision machine tools require δ_ω < 0.05% (machining error < 0.01 mm). max (ω) and min (ω) are the maximum and minimum rotational speeds, respectively.

The constraint of soft-start on machine accuracy improvement is reflected in the tool displacement deviation Δx, i.e., the following equation [48]

$$\Delta x={\displaystyle \frac{F_{\mathrm{shock}}}{K_s}}\cdot \cos \theta$$

(39)

where K_s is the machine rigidity, with a typical value of 10⁸ N/m for CNC machines; θ is the cutting angle, and soft-start makes F_shock ↓ → Δx ↓.

Table 5. Characteristics and benefits of motor soft-starting technology in different industries.

Application Sectors	Typical Equipment	Technological Needs	Key Benefits	Applicable Soft-Start Types
Power and Energy	Fans, pumps, and compressors	High-voltage, high-power, and grid-compatible	Reduces starting current and protects the grid	High-voltage magnetic control and solid-state soft-start
Petrochemical Industry	Process pumps and compressors	Explosion-proof and smooth starting	Avoid water hammer and operate safely	Magnetically controlled and explosion-proof soft-starters
Metallurgy and Mining	Crushers and ball mills	Heavy-duty starting and environmental resistance	High torque starting and equipment protection	High-current soft-start and magnetically controlled
Water Treatment and Municipal Engineering	Water pumps and ventilators	Anti-surge and energy-saving	Smooth control and energy-saving	Soft-start for pump control
Manufacturing and Machinery	Machine tools and conveyor belts	Precision control and low-noise	Reduced shock and increased precision	Electronic and brushless DC control

shows the characteristics and benefits of motor soft-starting technology in different industries.

In the traditional manufacturing industry, machine tools, conveyor belts, injection molding machines, and other equipment using soft-start technology can significantly reduce mechanical impact and improve processing accuracy and product consistency. Astronergy has a thyristor triggering anomaly protection function for soft-starters, providing highly reliable starting solutions for manufacturing applications. In the fields of packaging machinery, textile machinery, etc., multi-mode configuration has become a key feature in modern low-voltage starters. Recent commercial implementations allow users to select from distinct profiles (e.g., voltage ramp, current limit, or torque control) to adapt to diverse load requirements [48].

3.2. Technical Benefits of Soft Motor Starting

3.2.1. Energy-Saving and Emission Reduction Benefits

Energy-saving and emission reduction benefits are important values of soft-start technology. High-voltage solid-state soft-starting cabinet can effectively reduce motor energy consumption through precise control of starting current, reduce starting stress, and light-load energy-saving functions. The actual application data show that the use of soft-starting technology can make increase the motor system energy efficiency by 5~15%; in frequent start–stop scenarios, the energy-saving effect is more significant. In addition, magnetically controlled soft-starters can further improve the energy utilization rate during operation through the integration of power factor correction, reactive power compensation, and other advanced energy-saving technologies. From an environmental protection perspective, the reduction in energy consumption directly reduces carbon emissions, which is particularly critical in the context of increasing global climate change [49].

3.2.2. Equipment Protection and Life Extension

Another important value of soft-starting technology is reflected in equipment protection and life extension. Traditional direct-starting method produce excessive current and mechanical shock, such shock will significantly shorten the service life of motors and drive systems. Soft-starting technology gently regulates the starting process to avoid the above problems, thereby effectively protecting equipment and extending overall service life of the equipment [50]. According to the L10 standard, the bearing fatigue life L₁₀ is expressed as follows:

$$L_{10}={\left({\displaystyle \frac{C}{P}}\right)}^p\times {10}^6$$

(40)

where L₁₀ denotes the rated fatigue life, at which 90% of bearings can reach or exceed this life; C is the rated dynamic load corresponding to a bearing life of 1 million revolutions under constant load; p is the life of the index; for ball bearings, p = 3, and for roller bearings, p = 10/3; and P for the equivalent dynamic load, the actual load is converted to the equivalent radial load and can be expressed as follows:

$$P=X{F}_r+Y{F}_a$$

(41)

where F_r is the radial load; F_a is the axial load; X is the radial coefficient, obtained from the tables (determined by F_a/F_r and the contact angle), and X takes the value of the range of 0.4~1.5; and Y is the axial coefficient, obtained from the tables (determined by the type of bearings and F_a/C₀), and Y takes the value of the range of 0.6~2.5. Bearing life-related parameters and the impact of soft-starting on them are as follows: L₁₀ indicates 90% of bearing life (revolutions); the equivalent dynamic load P is calculated according to Equation (41); and C is the rated dynamic load. For impact loading, the equivalent dynamic load is P_direct = 2.8P_N for direct startup, and after soft startup it is reduced to P_soft = 1.2 P_N, where P_N is the rated load. The resulting life enhancement multiplier is expressed as follows:

$${\displaystyle \frac{{\displaystyle L_{10}^{soft}}}{{\displaystyle L_{10}^{direct}}}}={\left({\displaystyle \frac{2.8}{1.2}}\right)}^3\approx 15.6$$

(42)

By limiting starting current and torque, soft-start technology can reduce motor winding temperature rise by 30~50%, while greatly reducing the wear and tear of bearings and other mechanical parts, which is of great significance for ship propulsion systems in terms of vibration and noise reduction. Astronergy’s thyristor-triggered anomaly protection technology further improves system reliability and reduces maintenance frequency and cost. JISCO Group Hongxing Iron and Steel Co., Ltd.’s cold-running control technology solves the problem of long-term charged aging of the soft-starter itself. Overall, the use of soft-starting technology can extend the life of the motor system by 20% to 30% and reduce maintenance costs by 15% to 25%.

3.2.3. Intelligent Operation and Maintenance Advantages

The advantages of intelligent operation and maintenance are becoming more and more prominent with the progress of technology. Modern soft-starters are equipped with remote monitoring and communication functions, which allow users to monitor and operate devices in real time through smart devices or web platforms. Zhejiang Xinmai’s soft-starters support the configuration of operating parameters through communication interfaces, which realizes highly flexible control. These intelligent features not only simplify the equipment management process but also provide data support for predictive maintenance, helping to investigate potential failures and avoid unplanned downtime. Under the framework of industrial IoT, soft-starters act as intelligent terminals that are better able to work with other devices to further optimize energy efficiency and reliability across the entire production system [51].

Table 6. Technical suitability for motor soft-start applications.

Sector	Technical Pain Points	Prescription	Quantitative Benefits
Power and Energy	Grid shock (>6Irated)	High-voltage solid-state soft-start cabinet	Starting current ↓ 60%, protective relay life ↑ 3 times
Petrochemical Industry	Explosion-proof	Magnetic non-contact starting	ATEX-certified, accident rate ↓ 100%
Metallurgy and Mining	Heavy-duty mechanical stress	Segmented current control	Ball mill gearbox replacement cycle ↑ 40%
Water Treatment and Municipal Engineering	Water hammer damages pipes	Soft stop + closed-loop pressure	Pipeline maintenance costs ↓ 80%
Manufacturing and Machinery	Precision equipment shock	Microsecond torque control	CNC machine tool machining accuracy ↑ 22%

shows the motor soft-start application suitability. To quantitatively evaluate the reconstruction performance of the Inverse Distance Weighting (IDW) surrogate model, we conducted an error analysis using an independent test set derived from physical field simulations. Two standard metrics, Root Mean Square Error (RMSE) and Coefficient of Determination (R²), were employed to measure the deviation between the surrogate model predictions and ground truth. The model achieved an RMSE of 0.042 and an R² score of 0.985 on the test set. These results indicate a high degree of consistency between the surrogate model and actual physical fields, demonstrating that the IDW strategy effectively captures the spatial variations with sufficient accuracy for subsequent optimization tasks without incurring high computational costs.

3.3. Experimental Validation in Industrial Scenarios

To validate the proposed soft-start strategies, field tests were conducted across 12 distinct application scenarios, covering the power generation, mining, and petrochemical industries. The validation focused on key performance metrics, including peak starting current I_peak, startup duration (t_start), and mechanical vibration levels.

Experimental Setup: The validation setup involved high-voltage solid-state soft-starters (6 kV/10 kV class) and magnetic-controlled soft-starters deployed on site. Data were acquired using high-precision power quality analyzers (Fluke 435-II) and vibration sensors deployed on the motor drive train.

Representative Test Cases:

Table 7. Control strategies and application effectiveness verification under different typical loads.

Case ID	Load Type	Motor Specifications	Control Strategy	Measurement Conditions	Results
Case 1	Ball Mill	10 kV/2500 kW (High-Inertia)	Segmented Current Control	Heavy load start (J = 5000 kg × m2)	Ipeak: Reduced from 6.5 Ie to 2.8 Ie Vibration: Gearbox shock reduced by 40%
Case 2	Centrifugal Pump	6 kV/800 kW	Voltage Ramp + Soft Stop	Variable flow and closed valve	Pressure: Water hammer effect eliminated Stop Time: Extended from 5 s to 20 s (Linear)
Case 3	Induced Draft Fan	10 kV/5000 kW	Fuzzy Adaptive Control	Fan inertia fit (J = 1500 kg × m2)	Time: Optimized from 18 s to 12 s Thermal: Winding temp rise ↓ 15 °C
Case 4	Belt Conveyor	660 V/315 kW	Torque Control	Full load startup	Smoothness: Torque fluctuation (KT) < 5% Slippage: 0 events observed

presents a summary of four representative test cases selected from the twelve scenarios. The results compare the performance of the proposed intelligent soft-start with traditional direct-on-line (DOL) or reactor starting methods.

4. Trends and Challenges in Motor Soft-Start Technology

Table 8. Key directions for the future development of motor soft-start technology.

Direction of Development	Technical Characteristic	Potential Advantages	Key Challenges	Representative Companies/Technologies
Intelligent Adaptive Control	Machine learning algorithms and automatic parameter optimization [52]	Adapt to complex loads and reduce manual debugging	Algorithm complexity and real-time requirements	CoreMax Multi-Mode Control
Wide-Bandwidth Device Applications	SiC/GaN power devices	High efficiency and high-power density	High-cost and complex drive protection	Astronergy Power Electronics
System Integration	Multi-functional integration and modular design	Saves space and improves reliability	Thermal management and EMC Issues	Reynolds Bypass-Integrated Design
Special Field Applications	Highly reliable and miniaturized design	Expanding into emerging markets	Customization costs and certification requirements	Brushless DC Soft-Start

systematically elaborates the future development of motor soft-starting technology from four major development directions: intelligent adaptive control, wide-bandwidth device application, system integration, and special field applications, covering technical characteristics, potential advantages, major challenges, and representative enterprises/technologies. Motor soft-start technology has matured after years of development but still maintains a rapid evolutionary trend driven by emerging application needs and empowered by technological advances. Looking ahead, the technology will continue to break through in the direction of higher performance, greater intelligence, and more environmental friendliness; however, it is also necessary to deal with technical difficulties and market obstacles that remain to be resolved.

4.1. The Development Trend of Motor Soft-Start Technology

Table 9. Motor soft-start technology leaps in the course of the route process.

Orientations	Target 2025	Vision 2030	Key Supporting Technologies
Intelligent control	Load recognition accuracy > 90%	Autonomous generation of optimal starting curves	Federated learning + digital twins [53]
Wide-bandwidth devices [54]	SiC costs down to 2 × silicon-based	100 kW + all SiC solution popularized	Mass production of 8-inch SiC wafers
System integration	Half the cabinet volume	Chip-based soft-start modules	Three-dimensional packaging + liquid-cooled microchannels
Special applications	Space capsule prototype validation	Commercial deep space probe	Anti-radiation ASIC chip
Environmental standard	Full-life-cycle carbon traceability	Carbon neutral product certification	Blockchain carbon footprint platform

shows the motor soft-start technology across five directions—intelligent control, wide-bandwidth devices, system integration, special applications, and environmental standards—along with the 2025 target, the 2030 vision, and their respective key support technology leap route processes.

4.1.1. Intelligent and Adaptive Control Technology

Autonomous and Self-Evolving Control: Moving beyond current adaptive algorithms, the next generation of soft-starters will focus on autonomous learning. Future devices are expected to leverage digital twin technologies and deep reinforcement learning to generate optimal starting curves in real time, independent of manual parameter tuning. This represents a shift from configurable systems to fully self-driving motor control [55,56].

4.1.2. Wide-Bandgap Semiconductor Devices

The conduction loss of Si-based IGBT is P_condSi and the conduction loss SiC MOSFET is P_condSiC, respectively, can be expressed as follows:

$$P_{\mathrm{condSi}}={\displaystyle I_{\mathrm{rms}}^2}\cdot R_{\mathrm{ds}\left(\mathrm{on}\right)\mathrm{Si}}+U_{\mathrm{ce}0}\cdot I_{\mathrm{avg}}$$

(43)

$$P_{\mathrm{condSiC}}={\displaystyle I_{\mathrm{rms}}^2}\cdot R_{\mathrm{ds}\left(\mathrm{on}\right)\mathrm{SiC}}$$

(44)

where I_rms is the rms value of the current; I_avg is the average value of the current; R_{ds (on)} is the on-resistance, SiC ≈ 1/5 of the silicon-based IGBT (same withstand voltage); and U_ce0 is the IGBT threshold voltage.

The expression for the single switching energy E_sw can be expressed as follows:

$$E_{\mathrm{sw}}={\displaystyle \frac{1}{2}}{U}_{ds}\cdot I_d\cdot (t_r+t_f)+Q_{rr}\cdot U_{ds}$$

(45)

where t_r and t_f are the rise/fall times, respectively, and the rise/fall time of SiC is ≈1/5 (20 ns/100 ns) for SiC; and Q_rr is the reverse recovery charge, which is ≈0 (no reverse recovery).

The expression for the total switching loss P_sw can be expressed as follows:

$$P_{\mathrm{sw}}=(E_{\mathrm{on}}+E_{\mathrm{off}})\cdot f_{\mathrm{sw}}$$

(46)

where f_sw is the switching frequency, which can be boosted to 100 kHz for SiC (≤20 kHz for silicon-based).

Based on Equations (43)–(46), a comparative analysis was performed for a 100 kW motor drive system. Under the condition of a 20 kHz switching frequency, the negligible reverse recovery charge and faster switching speed of SiC devices allow for a theoretical reduction in total switching losses by approximately 80% compared with traditional Si-based IGBTs [57].

4.1.3. System Integration and Modular Design

System integration and modular design is an effective way to improve the reliability of soft-starters and reduce costs. Shanghai Reynolds Technology’s bypass-integrated design has demonstrated the advantages of integration. In the future, the device will be further integrated into motor protectors, frequency converters, and other equipment to form multi-functional integrated programs. Modular design allows users to flexibly configure functions (e.g., communication and energy return module) as required. The multiple starting modes and wide range of settings of the HSS soft-starter cabinet also reflect this flexibility. Integration also reduces external wiring and intermediate links, thereby improving reliability (such as the soft-start circuit of Shenzhen R&I Electric, which simplifies the control structure through an internal comparison circuit), while standardized interfaces and plug-and-play designs reduce the difficulty of installation and commissioning to shorten the project cycle.

4.1.4. New Energy and Special Applications

New energy sources and special applications offer new market opportunities for soft-start technology. In renewable energy systems such as photovoltaic and wind power, soft-start technology can optimize auxiliary equipment operation and improve overall system efficiency. Motor control systems in electric vehicles also draw on the experience of industrial soft-start technology to achieve smooth starting and energy recovery [4]. The demand for high-reliability motor control in special fields such as the aerospace and military industries will push soft-start technology to a higher standard. In addition, the miniaturization and specialization of soft-starters in the field of home appliances, medical equipment, and other areas of application are also showing a growing trend, such as the brushless DC motor soft-starting technology developed by Zhejiang Xinmai Technology Co., Ltd. for home appliances and other scenarios.

4.1.5. Energy Efficiency Standards and Environmental Requirements

Energy efficiency standards and environmental protection requirements will continue to drive innovation in soft-starter technology. Global motor energy efficiency regulations are becoming more stringent, requiring soft-starters not only to excel in the starting phase but also to pay attention to energy efficiency throughout the whole operation [58]. Furthermore, establishing precise methods and standards for quantitatively assessing the energy-saving and emission reduction benefits of soft-start technology is essential. This will assist users in accurately calculating the return on investment.

4.2. Technical Challenge

4.2.1. Technical Challenges and Bottlenecks

Technical challenges and bottlenecks still need to be overcome by the joint efforts of the industry. In high-voltage (e.g., above 10 kV) applications, series voltage equalization and synchronous triggering of power electronic devices are still technical difficulties [59]. Heat dissipation problem in high-power occasions also limits the miniaturization of the device. Although advanced heat dissipation technologies such as liquid cooling are being applied, their cost and reliability still need to be optimized. In terms of control algorithms, balancing the contradiction between starting time, current limitation, and torque response is still the focus of researchers’ attention. In addition, as the boundaries between soft-starters and other motor control equipment, such as frequency converters and servo systems, are gradually blurring, clear positioning and differentiation are strategic issues that enterprises need to think about.

4.2.2. Standardization and Ecosystem Building

Standardization and ecosystem building are crucial to the development of the soft-start industry. Currently, the technology lacks unified standards in the definition of terms, performance testing, and interface specifications, increasing the difficulty of user selection and supplier development costs. In the future, industry organizations need to take the lead in developing standards to promote technical exchange and market transparency. At the same time, ecosystem building for soft-start technology should not be ignored, covering design tools, simulation models, and training systems for installation and maintenance personnel. Shanghai Raynor Technology, Chint Electric, and other leading companies have been actively building their own technology ecosystems.

4.2.3. Cost and Market Acceptance

Cost and market acceptance are still real obstacles in the popularization process. Although the long-term economic benefits of soft-starting technology are significant, the initial investment is higher than that of traditional starting methods, which is a hindrance to price-sensitive small- and medium-sized enterprises [60]. Innovative products such as high-voltage dry-type magnetically controlled soft-starters are trying to crack this challenge by combining high performance with low cost. Market education is also crucial, as many users’ understanding of soft-start technology remains at the level of simple reduced-voltage starting, without understanding the diverse functions and comprehensive benefits of modern soft-starters. The industry therefore needs to help users fully recognize the value of soft-start technology through successful case sharing and energy efficiency assessment tools.

5. Summary and Outlook

This paper systematically reviews motor soft-start technology based on intelligent control and wide-bandgap semiconductors. Analysis of existing studies shows that replacing silicon-based devices with SiC/GaN devices can reduce switching losses by approximately 80% and support switching frequencies of up to 100 kHz. Furthermore, the integration of fuzzy adaptive control and multi-mode AI strategies has been shown to limit starting current to 1.5–3 I_e and keep torque-tracking error below 5%. The high-reliability and energy-saving benefits of these technologies are evidenced by their increasing adoption across various industrial sectors, including electric power and mining. In the future, soft-start technology will evolve in the direction of device miniaturization (SiC chip thickness < 100 μm), control autonomy (AI load recognition accuracy > 98%), and system integration (power density > 20 kW/kg), breaking through the bottlenecks related to high-voltage series voltage equalization and thermal management, while accelerating adaptation in special scenarios such as new energy, aviation, and aerospace, so as to contribute to the intelligence of industrial motor systems and the dual-carbon goals. Therefore, it will also accelerate the adaptation in new energy, aerospace, and other special scenarios, providing core support for the intelligence of industrial motor systems and the double-carbon goals.

While the proposed optimization strategy effectively balances parasitic parameters and thermal resistance, it is important to acknowledge that the electrical and thermal objectives are treated as relatively independent variables in the current iteration to maintain computational efficiency. However, in practical high-power applications, SiC MOSFETs exhibit a distinct positive temperature coefficient regarding on-resistance. Specifically, as the junction temperature rises, the on-resistance increases, leading to higher conduction losses, which in turn further elevates the temperature. Ignoring this strong electro-thermal coupling may result in an underestimation of the peak junction temperature under heavy load conditions. Future Improvement: To address this, our future work will aim to integrate a coupled electro-thermal solver where device material properties are dynamically updated based on the thermal field distribution during the iterative optimization process. This will ensure higher fidelity in reliability assessments, albeit at the cost of increased computational resources.