Design of a High-Voltage Insulation Resistance Detection System for Commercial Vehicles

Abstract

Based on the safety monitoring requirements of power batteries for new energy commercial vehicles, this study proposes a battery insulation detection method utilizing the bridge method in combination with existing insulation detection techniques. Through optimization and improvement of this methodology, an insulation detection system for new energy commercial vehicle power batteries was developed, along with a dynamic online calculation method for insulation resistance. Comprehensive experimental testing under multiple operational conditions demonstrates that the proposed system achieves a measurement accuracy with errors below 5% for insulation resistance. Furthermore, the system exhibits significant advantages in dynamic real-time monitoring capabilities and operational reliability.

1. Introduction

With the requirement of the dual-carbon strategy, the development of electric vehicles is necessary for sustainable transportation and upgrading the global automotive industry. China’s traditional automotive enterprises have overtaken passenger cars in terms of electrified technology. However, commercial vehicles are still following suit to start the electrification era. Most electric vehicles (EVs) utilize high-voltage lithium-ion batteries as their power source [1,2,3]. Given that high-voltage lithium-ion batteries are part of a high-voltage direct current (HVDC) system, the high-voltage insulation performance of the vehicle is of great significance to the safe operation of electric vehicles [4,5,6]. But electric commercial vehicles also use lithium batteries as their energy source, and the voltage platform is much higher (more than 800 V). This high-voltage level can be very dangerous to personal safety. The operating conditions and environments of electric vehicles, including high-temperature, high-humidity, and high-salinity fog environments, are highly complex [7,8,9,10,11]. Under these conditions, high-voltage electrical components may exhibit changes in high-voltage insulation performance, and in some cases, even insulation failure. Real-time monitoring of the changes in high-voltage insulation performance and responding to variations in insulation resistance are crucial for ensuring the safe operation of electric vehicles.

Traditional insulation detection methods can be categorized into two types: online and offline [12]. The voltmeter readings are only accessible in an offline manner [13,14]. The disadvantages of this technique include a limited voltage range, complex operation, and unstable voltage output.

Online insulation detection includes the use of balanced bridge methods, unbalanced bridge methods, and signal injection methods [15,16]. The balanced bridge system is a constant measurement system, which is not suitable for scenarios where insulation resistance (IR) decreases simultaneously on both sides. The unbalanced bridge method can overcome the shortcomings of the balanced bridge method but significantly increases the cost of the measurement circuit [17,18]. The disadvantages of the bridge methods mainly include low robustness, low detection accuracy, poor sensitivity, complex circuitry, and high cost. Shen et al. [19] improved upon the traditional bridge method by employing a high-precision instrumentation amplifier and optimizing the algorithm. Although this approach enhanced the detection accuracy and reliability, it lacks versatility and is not suitable for high-voltage platforms.

The signal injection method is a technique that involves injecting a low-frequency signal into the battery pack and calculating insulation based on the feedback signal [12]. This method is easy to implement and can achieve real online detection, gradually becoming the mainstream testing method for insulation resistance in high-voltage systems of electric vehicles. However, this method requires simultaneous computation of the Kalman filter algorithm and the Recursive Least Squares (RLS) algorithm, which increases computational complexity and limits its application in electric vehicles. Song et al. proposed an equivalent circuit model for insulation resistance measurement in high-voltage direct current (HVDC) transmission systems based on low-voltage and low-frequency signal injection. They employed the Extended Kalman Filter (EKF) algorithm to estimate the insulation resistance of HVDC systems, which can accurately estimate the insulation resistance [20]. However, the EKF algorithm is prone to significant round-off errors in nonlinear processes [21]. Tian et al. [12] proposed a detection method based on the Kalman filter, considering that insulation detectors are susceptible to noise interference. They employed the Recursive Least-Squares (RLS) method to address the issue of varying detection results with battery pack voltage. Zhen [22] introduced a low-frequency signal injection method and designed a least-squares algorithm with a variable forgetting factor to rapidly and accurately estimate insulation resistance and Y-capacitance values. This method features fast response (average response time of 3 s), high robustness, high estimation accuracy (Root Mean Square Error, RMSE, less than 0.012), and strong interference resistance, making it suitable for the promotion in electric vehicles. Bukya et al. [23] proposed a VFF-RLS algorithm based on Field-Programmable Gate Array (FPGA) technology. The VFF-RLS-based FPGA technique effectively reduces errors when dealing with changes in battery terminal voltage and resistance conditions and can quickly monitor variations in Insulation Resistance (IR). It is evident that real-time online monitoring and alerting of the insulation status of the power battery system in electric commercial vehicles are essential for ensuring the overall vehicle safety. To achieve this, it is crucial to perform dynamic, real-time, and effective detection of the insulation resistance value of the power battery system. An alarm should be triggered when the insulation resistance value falls outside the safety range, and the controller should take timely safety measures to protect personal safety.

This article provides a battery insulation detection algorithm based on the bridge-balancing methodology, with further optimizations to develop an insulation monitoring system and enhanced algorithm tailored for commercial new energy vehicle (NEV) traction battery systems. The proposed system demonstrates robust applicability in high-voltage commercial vehicle environments, characterized by exceptional universality and measurement accuracy exceeding industry benchmarks.

2. Design of Insulation Inspection System for Commercial Vehicles

Due to the complex system and high-voltage platform of electric commercial vehicles, the signal injection method suffers from significant dynamic detection errors and long response times. An article detection system and dynamic insulation detection algorithm for the battery system of electric commercial vehicles are proposed. These are based on the bridge method, incorporating the principles of the latest national standards and the improvement methods suggested by Shen Yongpeng et al. [19]. The insulation detection circuit is integrated into the Battery Management System (BMS) [24,25,26,27].

3. Battery Insulation Detection Method Based on Bridge Method

The battery insulation detection method based on the bridge method is shown in Figure 1. The detection steps are as follows:

(1) Two voltmeters with the same internal resistance are simultaneously in parallel across the two terminals of the REESS and between the electrical platform; the higher voltage is read s U₁, and the lower is read as U₁′, as shown in Figure 1b.

(2) After connecting a known unbalanced resistor R₀ to the end with a larger voltage, the voltages at this moment are then read as U₂ and U₂′, as shown in Figure 1c.

Figure 1. Battery Insulation Resistance Detection Principle Based on the Bridge Method. Notes: ① REESS (rechargeable electrical energy storage system) means a rechargeable energy storage system that can provide electrical energy; U_E means the electric potential of REESS, and I, I₁, and I₂ mean the current flowing through the corresponding circuit. ② R_i1 and R_i2 denote the insulation resistance between the REESS and the electrical platform, and R_i2 is the smaller of the resistance values, from which it is determined as the REESS insulation resistance R_i. ③ V₁ and V₁′ indicate the voltages across R_i1 and R_i2 in an ideal state. ④ U₁, U₁’, U₂, and U₂’ indicate the corresponding circuit voltmeter measurements. ⑤ r denotes the internal resistance of the corresponding voltmeter. ⑥ R₀ indicates unbalanced resistance.

Figure 1. Battery Insulation Resistance Detection Principle Based on the Bridge Method. Notes: ① REESS (rechargeable electrical energy storage system) means a rechargeable energy storage system that can provide electrical energy; U_E means the electric potential of REESS, and I, I₁, and I₂ mean the current flowing through the corresponding circuit. ② R_i1 and R_i2 denote the insulation resistance between the REESS and the electrical platform, and R_i2 is the smaller of the resistance values, from which it is determined as the REESS insulation resistance R_i. ③ V₁ and V₁′ indicate the voltages across R_i1 and R_i2 in an ideal state. ④ U₁, U₁’, U₂, and U₂’ indicate the corresponding circuit voltmeter measurements. ⑤ r denotes the internal resistance of the corresponding voltmeter. ⑥ R₀ indicates unbalanced resistance.

According to Ohm’s law and (b), the following can be obtained (“//” is the parallel symbol):

$${\displaystyle \frac{{U_1}^{\prime }}{U_1}}={\displaystyle \frac{R_{i2}//r•I_1}{R_{i1}//r•I_1}}$$

(1)

Metamorphosis must be:

$$R_{i1}//r=(R_{i2}//r){\displaystyle \frac{{U_1}^{\prime }}{U_1}}$$

(2)

According to Ohm’s law and (c),

$${\displaystyle \frac{{U_2}^{\prime }}{U_2}}={\displaystyle \frac{R_{i2}//r•I_2}{R_{i1}//r//R_0•I_2}}$$

(3)

Substituting Equation (2) into Equation (3) yields:

$${\displaystyle \frac{R_{i2}•r}{R_{i2}+r}}=R_0({\displaystyle \frac{{U_2}^{\prime }}{U_2}}-{\displaystyle \frac{{U_1}^{\prime }}{U_1}})$$

(4)

Rearranging the equation yields:

$$R_{i2}={\displaystyle \frac{1}{\frac{1}{R_0(\frac{{U_2}^{\prime }}{U_2}-\frac{{U_1}^{\prime }}{U_1})}-\frac{1}{r}}}$$

(5)

The values of U₁, U₁′, U₂, and U₂′ can be obtained through voltage measurement equipment. The values of R₀ and r are known. The value of R_i2 can be calculated using Equation (5), which represents the insulation resistance value R_i of the entire vehicle.

But there are two problems with the method:

1. Detection accuracy is determined by the detection accuracy of U₁, U₁′, U₂, and U₂′. Due to the influence of the distribution capacitance, the higher the total battery voltage is, the longer the stability of U₂ and U₂′ after incorporating the imbalanced resistor R₀, and the greater the error that exists in the detection, which results in the greater the error in the calculated insulation resistance value.

2. When simultaneous insulation faults occur between the REESS ends and the electrical platform, only the smaller of the insulation resistances can be calculated, and it is not possible to calculate R_i1 and R_i2 at the same time.

Verification Standard for Insulation Monitoring Function of Electric Vehicles

R_x is defined as the target value for the insulation fault alarm, U_REESS is the current total voltage of the battery pack, and R_i is the current insulation resistance value of the whole vehicle. If the minimum insulation resistance value is required to be 100 Ω/V, R_x is defined in Equation (6) in the range for audible and visual alarms.

$${\displaystyle \frac{1}{\frac{1}{95U_{REESS}}-\frac{1}{R_i}}}\le R_x<{\displaystyle \frac{1}{\frac{1}{100U_{REESS}}-\frac{1}{R_i}}}$$

(6)

If the minimum required insulation resistance value is 500 Ω/V, then, if the value of R_x falls within the range specified by Equation (7), an audio-visual alarm will be triggered.

$${\displaystyle \frac{1}{\frac{1}{475U_{REESS}}-\frac{1}{R_i}}}\le R_x<{\displaystyle \frac{1}{\frac{1}{500U_{REESS}}-\frac{1}{R_i}}}$$

(7)

In this paper, the voltage platform of the commercial vehicle power battery system is 800 V, i.e., U_REESS = 800 V. Under normal circumstances, R_i is 10 MΩ or more, allowing 1/R_i to be ignored; the actual design scheme will generally design a multi-level warning mechanism, as detailed in

Table 1. Insulation warning mechanisms.

Fault Level	Resistance Range	Alarm Action
Class II	76 kΩ ≤ Rx < 80 kΩ	audible and visual alarm
Class I	380 kΩ ≤ Rx < 400 kΩ	insulation warning
/	Rx ≥ 400 kΩ	normal conditions, no alarm

.

4. Power Battery Insulation Detection Circuit Design and Estimation Algorithm for Commercial Vehicles

Aiming at the problems existing in the battery insulation detection method based on the bridge method, according to the special characteristics of commercial vehicles’ working conditions, voltage platforms, and vehicle systems, we reviewed the existing mainstream insulation detection methods, and based on this method, we proposed an insulation resistance detection circuit for power batteries in commercial vehicles, as shown in Figure 2.

The circuit uses the vehicle ground as a reference and consists of five modules: power battery pack (including the Y capacitance of the whole vehicle), insulation resistance, unbalanced bridge arm, sampling bridge arm, and MCU. Among these, the unbalanced bridge arm, sampling bridge arm, and MCU modules constitute the hardware circuit components for the insulation resistance detection of the commercial vehicle’s power battery. The unbalanced bridge arm, sampling bridge arm module consists of four optical MOS (MOSFET) switches, a pair made up of an imbalanced bridge arm and a resistance divider sampling arm, where R₀ = R_0p = R_0n, R_a = R_ap = R_an, and R_b = R_bp = R_bn. When all four switches are open, the insulation detection circuit can be disconnected from the positive and negative busbars, thereby preventing the insulation detection circuit from affecting other circuits. The opening and closing of switches S₃ and S₄ are controlled through software settings, thereby controlling the termination and initiation of insulation detection. The opening and closing states of switches S₁ and S₂ are set to control the busbar end connected to the unbalanced resistor R₀. The sampling bridge arm acquires V_p and V_n in different states by the resistor voltage divider. The MCU (Microcontroller Unit) acquires the AD value of the voltage signals, then performs signal processing, conversion, and calculates the insulation resistance value, and then makes corresponding alarm actions according to the fault mechanism. The method requires only three steps to calculate the insulation resistance value.

1. Positive unbalanced voltage acquisition

Close S₁, S₃, and S₄; disconnect S₂ to obtain U_p1, U_n1, V_p1, and V_n1; the equivalent circuit is shown in Figure 3a, where R_c = R_ap + R_bp = R_an + R_bn, R_p′ = R_p//R_c, R_n′ = R_n//R_c, and according to Ohm’s law, the following can be obtained:

$${\displaystyle \frac{U_{n1}}{U_{p1}}}={\displaystyle \frac{\frac{V_{n1}(R_{an}+R_{bn})}{R_{an}}•I_1}{\frac{V_{P1}(R_{ap}+R_{bp})}{R_{ap}}•I_1}}$$

(8)

$${\displaystyle \frac{U_{n1}}{U_{p1}}}={R_n}^{\prime }•({\displaystyle \frac{1}{R_0}}+{\displaystyle \frac{1}{{R_p}^{\prime }}})$$

(9)

2. According to Equations (9) and (10), the following can be obtained:

$${R_n}^{\prime }={\displaystyle \frac{V_{n1}}{V_{p1}}}•({\displaystyle \frac{R_0•{R_p}^{\prime }}{R_0+{R_p}^{\prime }}})$$

(10)

3. Negative unbalanced voltage acquisition

Close S₂, S₃, and S₄; disconnect S₁ to obtain U_p2, U_n2, V_p2, and V_n2; the equivalent circuit is shown in Figure 3b, and the following can be obtained:

Figure 3. Equivalent circuit diagram of power battery insulation detection circuit for commercial vehicles.

Figure 3. Equivalent circuit diagram of power battery insulation detection circuit for commercial vehicles.

$${\displaystyle \frac{V_{p2}}{V_{n2}}}={R_p}^{\prime }({\displaystyle \frac{1}{R_0}}+{\displaystyle \frac{1}{{R_n}^{\prime }}})$$

(11)

Rearranging the equation yields:

$${R_p}^{\prime }={\displaystyle \frac{V_{p2}}{V_{n2}}}•({\displaystyle \frac{R_0•{R_n}^{\prime }}{R_0+{R_n}^{\prime }}})$$

(12)

4. Estimation of insulation resistance

Substituting Equation (10) into Equation (11) yields:

$${R_p}^{\prime }={\displaystyle \frac{R_0•(V_{p2}•V_{n1}-V_{p1}•V_{n2})}{V_{n2}•(V_{n1}+V_{p1})}}$$

(13)

By replacing Formula Equation (12) and substituting Formula Equation (9), we obtain:

$${R_n}^{\prime }={\displaystyle \frac{R_0•(V_{p2}•V_{n1}-V_{p1}•V_{n2})}{V_{p1}•(V_{n2}+V_{p2})}}$$

(14)

Based on R_c = R_ap + R_bp = R_an + R_bn, R_p ′ = R_p//R_c, R_n ′ = R_n//R_c, we obtain:

$$R_p={\displaystyle \frac{1}{\frac{(V_{n1}+V_{p1})•V_{n2}}{R_0•(V_{p2}•V_{n1}-V_{p1}•V_{n2})}-\frac{1}{R_{ap}+R_{bp}}}}$$

(15)

$$R_n={\displaystyle \frac{1}{\frac{(V_{n2}+V_{p2})•V_{p1}}{R_0•(V_{p2}•V_{n1}-V_{p1}•V_{n2})}-\frac{1}{R_{an}+R_{bn}}}}$$

(16)

5. Software Workflow of Insulation Detection Algorithm

5.1. Software Flow of Battery Insulation Detection Algorithm Based on Bridge Method

The circuit corresponding to the battery insulation detection method based on the bridge method is shown in Figure 2, and the detailed software flowchart is shown in Figure 4; after the insulation detection starts, the process starts by closing switches S₃ and S₄ to calculate U₁ and U₁′. Next, close the switch corresponding to the larger value U₁ end (S₁ or S₂), then calculate U₂ and U₂′, and finally, calculate the insulation resistance value Ri according to Formula (5), where r = R_ap + R_bp = R_an + R_bn.

5.2. Dynamic Insulation Detection Algorithm Software Flow for Commercial Vehicles

The dynamic insulation detection algorithm for commercial vehicles designed in this paper can simultaneously collect the insulation resistance values between the two ends of REESS and the electric platform. It improves sampling accuracy under the dynamic working conditions (with Y capacitance) of the whole vehicle by cyclically collecting the AD values on R_ap and R_an, and calculating the corresponding U_p and U_n. The algorithm also assesses the three consecutive voltage deviation values of U_p and U_n. The workflow diagram of dynamic insulation detection is shown in Figure 5.

The detection process includes the following steps:

1. Initialization;

2. Close the positive unbalanced switch S₁ and the positive-negative sampling bridge arm S₃ and S₄;

3. Sample the AD values corresponding to V_p1 and V_n1 within 15 s and calculate V_p1, V_n1, U_p1, and U_n1;

4. When the deviation of U_p1 and U_n1 is ≤1 V three times consecutively, take the last V_p1 and V_n1 corresponding to U_p1 and U_n1 as the final value;

5. If U_p1 and U_n1 do not satisfy the three consecutive deviations of ≤1 V in 15 s, then the maximal insulation value and the insulation fault will be reported;

6. Close the negative unbalanced switch S₂ and the positive-negative sampling bridge arm S₃ and S₄; cyclically sample the AD values corresponding to V_p2 and V_n2 within 15 s and calculate V_p2, V_n2, U_p2, and U_n2; when the deviation of U_p2 and U_n2 is ≤1 V three times consecutively, take the last V_p2 and V_n2 corresponding to U_p2 and U_n2 as the final value; if U_p2 and U_n2 do not satisfy the three consecutive deviations of ≤1 V in 15 s, then the maximal insulation value and the insulation fault will be reported;

7. The final value of V_p1, V_n1, V_p2, V_n2, R_ap, R_bp, R_an, and R_bn is substituted into Equations (15) and (16) to calculate the insulation resistance of the positive and negative bus to the chassis. The detailed flow chart is shown in Figure 6.

As can be seen in Figure 6, there are two differences between the adaptive dynamic insulation inspection process for commercial vehicles and the battery insulation inspection process based on the bridge method: ① Closing the positive end unbalanced arm and the negative end unbalanced arm, respectively, which is done in order to calculate the insulation resistances of the busbar on the chassis at both the positive and negative ends, instead of only calculating the insulation resistance of the busbar on the chassis at both the positive and negative ends of the chassis at the smaller one. ② A sampling stability determination mechanism was implemented, requiring the system to meet predefined stability criteria within 15 s before proceeding to subsequent processes. Failure to satisfy these conditions triggers an insulation sampling fault alarm, aiming to enhance measurement accuracy by ensuring signal integrity.

6. Validation of Dynamic Insulation Test Methods for Commercial Vehicles

To verify the accuracy of the dynamic insulation detection method for new energy commercial vehicles and compare it with the battery insulation detection method based on the bridge method, a test bench (as shown in Figure 7) was established for validation. Under the same experimental conditions and hardware circuitry, tests were conducted using both the battery insulation detection method based on the bridge method and the dynamic insulation detection method for new energy commercial vehicles. The experiment uses a high-voltage power supply to simulate the battery pack voltage, which was set to 800 V. The developed circuit was integrated into the main controller of the battery management, and the programmable power supply was used to supply power to the main control board system; the power supply voltage was set to 24 V. In the experiment, the adjustable resistor box was used to connect resistors with different resistance values in parallel between the battery terminals and the vehicle ground (the vehicle body), thereby simulating different insulation resistance scenarios. Meanwhile, research revealed that the actual Y capacitance of the electric commercial vehicle has two kinds: 470 nf and 690 nf (formed by the parallel 470 nf and 220 nf). The experiment simulates the vehicle’s Y capacitance by incorporating 470 nf and 690 nf safety capacitors into the test circuit in parallel, respectively.

The test is carried out in four stages, corresponding to four different test states. State 1 is to access the simulated adjustable insulation resistance between the vehicle ground and the positive battery terminal or the vehicle ground and the negative battery terminal; state 2 is to access the simulated insulation resistance between the vehicle ground and the positive battery terminal, the vehicle ground and the negative battery terminal at one end, and a fixed insulation resistance at the other end; state 3 is to access 470 nf Y capacitance between state 3 is based on state 1, and at the same time, between the vehicle ground and the battery positive end, the vehicle ground and the battery negative end each has access to the 470 nf Y capacitor; state 4 is based on state 1, and at the same time, between the vehicle ground and the battery positive end, the vehicle ground and the battery negative end each has access to the 690 nf Y capacitor; the specific operation steps are shown below:

State 1: At first, variable resistor R was connected between the positive terminal of the battery and the vehicle ground; the negative terminal of the battery was left disconnected; measurements were run. Then, variable resistor R was connected between the negative terminal of the battery and the vehicle ground; the positive terminal of the battery was left disconnected; again, measurements were run. The insulation resistance value and the error data of the test are shown in

Table 2. State 1 test data.

Detection Methods	R Setting Value (kΩ)	Rp Measured Value (kΩ)	Rp Error	Rn Measured Value (kΩ)	Rn Error
Dynamic Insulation Test Method with unbalanced bridge	50	49	−2.00%	46	−8.00%
	100	99	−1.00%	91	−9.00%
	300	294	−2.00%	287	−4.33%
	500	484	−3.20%	472	−5.60%
	760	727	−4.34%	714	−6.05%
Dynamic Insulation Test Method for Electric Commercial Vehicles	50	50	0.00%	50	0.00%
	100	100	0.00%	100	0.00%
	300	301	0.33%	297	−1.00%
	500	500	0.00%	491	−1.80%
	760	761	0.13%	749	−1.45%

.

State 2: Firstly, variable resistor R was connected between the positive terminal of the battery and the vehicle ground, and a fixed resistor between the negative end of the battery and the vehicle ground, with a resistance value of 1.352 MΩ. Then, variable resistor R was connected between the negative terminal of the battery and the vehicle ground, and a fixed resistor between the positive end of the battery and the vehicle ground, with a resistance value of 1.352 MΩ. The insulating resistance values and the error data of the test are shown in

Table 3. State 2 test data.

Sensing Methodologies	Rp and Rn Setting Value (kΩ)	Rp Measured Value (kΩ)	Rn Measured Value (kΩ)	Rp Error	Rn Error
Dynamic Insulation Test Method with unbalanced bridge	50, 1352	49	/	−2.00%	/
	100, 1352	100	/	0.00%	/
	300, 1352	293	/	−2.33%	/
	500, 1352	478	/	−4.40%	/
	760, 1352	717	/	−5.66%	/
	1352, 50	/	50	/	0.00%
	1352, 100	/	97	/	−3.00%
	1352, 300	/	287	/	−4.33%
	1352, 500	/	473	/	−5.40%
	1352, 760	/	713	/	−6.18%
Dynamic Insulation Test Method for Electric Commercial Vehicles	50, 1352	49	1353	−2.00%	0.07%
	100, 1352	100	1357	0.00%	0.37%
	300, 1352	300	1353	0.00%	0.07%
	500, 1352	494	1345	−1.20%	−0.52%
	760, 1352	755	1335	−0.66%	−1.26%
	1352, 50	1351	48	−0.07%	−4.00%
	1352, 100	1322	98	−2.22%	−2.00%
	1352, 300	1343	295	−0.67%	−1.67%
	1352, 500	1344	488	−0.59%	−2.40%
	1352, 760	1347	746	−0.37%	−1.84%

Note: The bridge-based battery insulation detection methodology exhibits a critical limitation: it exclusively quantifies the insulation resistance at the terminal with lower resistivity, while failing to accurately measure the counterpart at the higher-resistance terminal due to inherent systemic constraints, according to the columns with “/”.

.

State 3: A 470 nf safety capacitor was connected in parallel between the positive end of the battery and the vehicle ground, as well as between the negative end of the battery and the vehicle ground, to simulate the Y capacitance of the whole vehicle. A variable resistor was sequentially connected between the positive end of the battery and the vehicle ground, the negative end of the battery and the vehicle ground, with the other end suspended in the air. The tested insulation resistance value and error data are shown in

Table 4. State 3 test data.

Detection Methods	R Setting Value (kΩ)	Rp Measured Value (kΩ)	Rp Error	Rn Measured Value (kΩ)	Rn Error
Dynamic Insulation Test Method with unbalanced bridge	50	48	−4.00%	42	−16.00%
	100	99	−1.00%	91	−9.00%
	300	284	−5.33%	245	−18.33%
	500	440	−12.00%	350	−30.00%
	760	597	−21.45%	495	−34.87%
Dynamic Insulation Test Method for Electric Commercial Vehicles	50	50	0.00%	49	−2.00%
	100	99	−1.00%	98	−2.00%
	300	297	−1.00%	294	−2.00%
	500	493	−1.40%	481	−3.80%
	760	743	−2.24%	733	−3.55%

.

State 4: To simulate the vehicle Y capacitor, a 690 nf safety capacitor was connected in parallel between the positive terminal of the battery and the vehicle ground, and between the negative terminal of the battery and the vehicle ground. A variable resistor was sequentially connected between the positive terminal of the battery and the vehicle ground, and between the negative terminal of the battery and the vehicle ground, while no resistor was connected at the other end. The measured insulation resistance values and error data are shown in

Table 5. State 4 test data.

Detection Methods	R setting Value (kΩ)	Rp Measured Value (kΩ)	Rp Error	Rn Measured Value (kΩ)	Rn Error
Dynamic Insulation Test Method with unbalanced bridge	50	49	−2.00%	41	−18.00%
	100	98	−2.00%	79	−21.00%
	300	262	−12.67%	191	−36.33%
	500	366	−26.80%	266	−46.80%
	760	454	−40.26%	310	−59.21%
Dynamic Insulation Test Method for Electric Commercial Vehicles	50	49	−2.00%	49	−2.00%
	100	95	−5.00%	97	−3.00%
	300	296	−1.33%	293	−2.33%
	500	484	−3.20%	476	−4.80%
	760	732	−3.68%	723	−4.87%

.

The results from the four sets of data demonstrate that, under identical hardware circuit and test environment conditions, the dynamic insulation detection method for new energy commercial vehicles is capable of not only calculating the insulation resistance between the positive or negative terminal of the battery and the vehicle ground, but also, in the event of simultaneous insulation abnormalities between the positive terminal of the battery, the negative terminal of the battery, and the vehicle ground, it can accurately calculate the insulation resistance among these components. Furthermore, an additional advantage of this dynamic insulation detection method is its optimization of the sampling method and algorithm in accordance with the characteristics of the commercial vehicle’s high-voltage platform and large Y capacitance, thereby enhancing the estimation accuracy. Verification has confirmed that the detection error of this method remains within 5% across various operating conditions. The application of this detection method significantly improves the insulation detection accuracy of commercial vehicles under the four defined operational scenarios, reduces the likelihood of false alarms for insulation faults, and thereby better ensures the personal safety of passengers.

In summary, the dynamic insulation testing method for electric commercial vehicles has better testing accuracy than the battery insulation detection method based on the bridge method, and compared with other existing insulation testing methods for electric passenger vehicles, this testing method can better meet the insulation testing needs of commercial vehicles.

7. Conclusions

The paper introduces two insulation detection methods and proposes a new dynamic insulation detection method applicable to commercial electric vehicles with high-voltage platforms, which can dynamically calculate the equivalent insulation resistance value of the positive and negative terminals of the high-voltage battery to the ground of the vehicle body in real time. The testing proves that the detection accuracy is within 5% under various working conditions, and at the same time, based on different insulation threshold values, insulation fault early warning can be achieved, which enables both single-end insulation fault alarms and dual-end insulation fault alarms. This method has been effectively implemented in the 800 V power battery system.

At the same time, the battery management system controller applying this testing method has already completed batch mounting on the new energy commercial vehicles of famous brands in China, which can carry out real-time and accurate testing of the insulation resistance value of the whole vehicle under various working conditions and provide fault alarms when there is an insulation risk in the vehicle to ensure the personal safety of the vehicle occupants. For traction battery systems operating within 800–1000 V platforms, the dynamic insulation monitoring methodology achieves accurate detection and early warning when insulation resistance resides in the range of 0–600 kΩ (equivalent to 600 Ω/V under the rated voltage).