Device and method for determining injection quality determination in hydrogen injector

Abstract

The controller 100 measures a real-time hydrogen pressure in the hydrogen rail 106, the hydrogen rail 106 being fluidly connected to the HGI 110 and the hydrogen pressure being measured before and after injection in the combustion chamber 112, characterized in that the controller 100 calculates a pressure drop in the hydrogen rail 106 and estimates the injected hydrogen mass using a trained machine learning model 108 by using the calculated pressure drop. The controller 100 determines an incremental mass by calculating a difference between the estimated hydrogen mass and a setpoint mass according to operating conditions of the engine, and determines an injection mass deviation of the HGI 110 by comparing the incremental mass to a threshold.

Description

Technical Field

[0001] The present invention relates to an apparatus and a method for determining injection quality deviations in a hydrogen gas injector (HGI). Background Art

[0002] In current practice, fuel flows from a storage tank to the combustion chamber. Hydrogen fuel is stored at high pressure in the tank, the pressure is regulated using a hydrogen injection pressure regulator, and the reduced-pressure gas is accumulated in a hydrogen rail. The hydrogen fuel is then injected into the combustion chamber via a hydrogen injector for accurate injection that facilitates proper combustion. However, the injected mass of hydrogen into the combustion chamber is not always equal to the setpoint mass, which slightly impacts the expected behavior of the system. Due to temperature constraints within the combustion chamber, calculating the precise amount of hydrogen injected into the combustion chamber using hardware / measurement equipment is difficult. Consequently, there is a strong incentive to determine injection mass deviations in hydrogen injectors.

[0003] According to the prior art US2018100449, a device for operating a gas fuel injector in an internal combustion engine. The device includes: a mass flow sensor that generates a signal representing the mass flow rate of gas fuel in a supply pipe in the engine. A controller is connected to the injector, and the mass flow sensor is programmed to actuate the injector to introduce gas fuel into the engine, which determines the actual mass flow rate of the gas fuel based on the signal representing the mass flow rate. It also calculates the difference between the actual mass flow rate and the expected mass flow rate. Based on this, when the absolute value of the difference is greater than a predetermined value, at least one of the on-time of the gas fuel injector and the magnitude of the injector activation signal is adjusted by a corresponding amount based on the difference. BRIEF DESCRIPTION OF THE DRAWINGS

[0004] The embodiments of the present disclosure are described with reference to the following drawings: Figure 1 FIGURE 1 illustrates a block diagram of an apparatus for determining injection quality deviation in a hydrogen gas injector (HGI) according to an embodiment of the present invention; Figure 2 A flow chart illustrating a method for determining injection quality deviation in a hydrogen gas injector (HGI) according to the present invention is shown, and Figure 3 Illustrated is a graph illustrating the relationship between hydrogen rail pressure and injector energization in accordance with the present invention. DETAILED DESCRIPTION

[0005] Figure 1A block diagram of an apparatus according to an embodiment of the present invention is shown, comprising a controller 100 for determining injection quality / quantity deviation in a hydrogen gas injector (HGI) 110. The HGI 110 is shown as part of a vehicle, but is not limited thereto. The vehicle includes a hydrogen tank 102, which stores high-pressure hydrogen gas, integrated with a temperature sensor. From the hydrogen tank 102, the hydrogen gas passes through a hydrogen injector pressure regulator 104 to a hydrogen rail 106. The gas flow path between the hydrogen tank 102 and the hydrogen injector pressure regulator 104 includes various filters and sensors for regulating the pressure of the hydrogen gas; however, these filters and sensors are not shown in the figure because they are not relevant to the explanation of the present invention and are also prior art. The hydrogen injector pressure regulator 104 includes a pressure regulator and a pressure sensor, which regulate the pressure of the hydrogen gas. From the hydrogen injector pressure regulator 104, the hydrogen gas enters a hydrogen rail 106, which is integrated with a pressure sensor 114, and ultimately, through the hydrogen injector 110, enters a combustion chamber 112 of the vehicle's engine.

[0006] According to an embodiment of the present invention, the gas accumulated / stored in the hydrogen rail 106 is at a high pressure compared to atmospheric pressure. The pressure at the hydrogen rail 106 is measured by an integrated pressure sensor 114. However, once the gas is injected into the combustion chamber 112, the pressure in the hydrogen rail 106 drops, and this pressure drop is used as one of the parameters to calculate the hydrogen injected into the combustion chamber 112 using an offline trained machine learning model 108. The machine learning model 108 is trained on engine speed, rail pressure before hydrogen injection, the drop in rail pressure due to injection, hydrogen rail 106 temperature, injector energization duration, engine torque, and mass flow through the hydrogen injector pressure regulator 104, among other factors. However, as injector energization varies across these parameters, these parameters are independent to achieve a certain vehicle speed based on driver demand.

[0007] According to an embodiment of the present invention, data is collected from an engine or test bench to reproduce / represent an injection scenario. The collected data is used to train a machine learning model 108 offline and port the machine learning model 108 to an engine control unit (ECU) or in the cloud. In a real-world scenario, once the vehicle is operating on the road, the trained machine learning model 108 reads the rail pressure across the injection and the engine speed, and uses the injector energization duration, engine torque, and mass flow through the hydrogen injector pressure regulator 104 to estimate the actual fuel mass injected into the combustion chamber 112.

[0008] According to an embodiment of the present invention, the incremental mass is determined by calculating the difference between the estimated mass of hydrogen injected into the combustion chamber 112 from the machine learning model and the setpoint mass according to the operating conditions of the engine. The determined incremental mass of hydrogen is compared with a predefined threshold to determine the drift or deviation in the injection mass through the hydrogen injector 110. The incremental mass is a drift or error that is used for further action, such as correction functions in the software or replacement of the hydrogen injector 110.

[0009] According to an embodiment of the present invention, the incremental mass value is compared to a threshold value, and based on the comparison of whether the incremental mass is within an allowable range, the ECU corrects the HGI 110 tolerance. If the incremental mass is outside the allowable range, correction is not possible and the hydrogen injector 110 needs to be replaced. For example, in a scenario where the incremental mass of hydrogen is above 80 percent of the threshold allowable range, it is an indication to change the hydrogen injector 110. In another scenario where the incremental mass of hydrogen is within the 80 percent limit of the allowable threshold range, the hydrogen injector 110 may be corrected.

[0010] According to an embodiment of the present invention, HGI 110 needs to be replaced or corrected, so that decisions can be made quickly and further damage can be avoided. The determined jet quality deviation of HGI 110 is transmitted to the cloud to update the status of HGI 110. From the cloud, the updated information is used by the OEM or end user, ensuring that both the end user and the OEM have the same understanding of the component status.

[0011] According to an embodiment of the present invention, controller 100 is at least one selected from the group of the device including smart phone, computer and cloud.Controller 100 is a controller including the following: input interface, output interface with pin or port, memory element such as random access memory (RAM) and / or read-only memory (ROM), analog-to-digital converter (ADC) and digital-to-analog converter (DAC), clock, timer, counter, and at least one processor (capable of realizing machine learning) connected to each other and to other components through a communication bus channel.Memory element (not shown) is pre-stored with logic or instruction or program or application or module / model and / or threshold value / range, reference value, predefined / predetermined criteria / condition, list, knowledge source that are accessed by the at least one processor according to defined routine.The internal components of controller 100 are not explained owing to being prior art, and it is not necessary to be understood in a limiting manner. The controller 100 may also include a communication unit for communicating via wireless or wired means, such as a transceiver, and the wireless or wired means are such as Global System for Mobile Communications (GSM), 3G, 4G, 5G, Wi-Fi, Bluetooth, Ethernet, serial network, etc. The controller 100 may be implemented as a system-in-package (SiP) or a system-on-chip (SOC), or any other known type. Examples of the controller 100 include, but are not limited to, a microcontroller, a microprocessor, a microcomputer, an electronic control unit (ECU), etc.

[0012] According to the present invention, the operation of the controller 100 is envisioned. The controller 100 uses the trained machine learning model 108 to estimate the mass of hydrogen (A) injected into the combustion chamber 112. To calculate the injected hydrogen mass, the pressure drop in the hydrogen rail 106 is used by the controller 100. Using the injected hydrogen mass A, the controller 100 calculates an incremental mass C as the difference between the estimated injected hydrogen mass A and the setpoint mass B according to the operating conditions of the engine. Finally, the deviation in the injected hydrogen mass D of the HGI 110 is determined by comparing the incremental mass C with a threshold value E. Once D is determined, it is transmitted to the cloud to update the status of the HGI 110, and based on the updated information, decisions can be taken accordingly.

[0013] Figure 2A flow chart of a method for determining injection mass deviation in a hydrogen gas injector (HGI) according to the present invention is illustrated. The method includes a plurality of steps, wherein step 202 includes measuring, by a controller, the real-time hydrogen pressure in a hydrogen rail 106 before and after injection, the hydrogen rail 106 being fluidically connected to the hydrogen gas injector 110. Step 204 includes calculating, by the controller, a pressure drop in the hydrogen rail 106; and estimating, using the calculated pressure drop, the mass of hydrogen injected into the combustion chamber 112 using the trained machine learning model 108. Step 206 includes determining, by the controller, an incremental mass using a difference between the estimated hydrogen mass and a setpoint mass according to operating conditions of the vehicle's engine. Step 208 includes determining, by the controller, an injection mass deviation based on a comparison between the determined incremental mass and a threshold value.

[0014] According to the method, step 204 further includes estimating the mass of hydrogen injected using the trained machine learning model 108, wherein the machine learning model 108 is trained using parameters including engine speed, fuel rail pressure before injection, fuel rail pressure drop due to injection, fuel rail temperature, injector energization, engine torque, and mass flow through the hydrogen injector pressure regulator 104. Once the machine learning model 108 is trained, it is deployed in an ECU or in the cloud. While the vehicle is operating on the road, the trained model reads the rail pressure across the injection and engine speed, and uses the injector energization duration, engine torque, and mass flow through the hydrogen injector pressure regulator 104 to calculate the actual mass of hydrogen injected into the combustion chamber 112.

[0015] Figure 3 Illustrated is a graph illustrating the relationship between hydrogen rail pressure and injector energization in accordance with the present invention. Figure 3 The rail pressure drop (P1 to P2) once hydrogen is injected into the combustion chamber 112 is explained. Here, P1 and P2 are the pressures measured before and after the hydrogen injector 110 is energized. Once the curve for the hydrogen injector 110 is energized, the curve for the hydrogen rail 110 pressure drops and a pressure difference (P1-P2) is present. The pressure difference (P1-P2) gives the amount of hydrogen injected into the combustion chamber 112. However, temperature dynamics affect this calculation, and therefore, a trained machine learning model 108 is used to determine the amount of hydrogen injected, which uses all relevant parameters such as engine speed, injector energization duration, engine torque, and mass flow through the hydrogen injector pressure regulator 104.

[0016] According to the present invention, a controller and method for determining injection quality deviations in an HGI 110 are disclosed. By determining deviations in injected quality, correction of the HGI is achieved. The determined deviations are transmitted to the cloud to update the status of the HGI 110. From the cloud, the updated information can be used by the OEM or end user, ensuring that both the end user and the OEM have the same understanding of component status. This information facilitates timely decision-making.

[0017] It should be understood that the embodiments explained in the above description are merely illustrative and do not limit the scope of the present invention. Many such embodiments and other modifications and variations of the embodiments explained in the description are contemplated. The scope of the present invention is limited only by the scope of the claims.

Claims

1. An apparatus for determining injection quality deviation in a hydrogen gas injector (HGI) (110), the apparatus comprising a controller (100), the controller (100) being configured to: a. Measuring real-time hydrogen pressure in a hydrogen rail (106) fluidly connected to the HGI (110), wherein the hydrogen pressure is measured before and after injection into a combustion chamber (112) of an engine, characterized in that: b. calculating the pressure drop in the hydrogen rail (106) and estimating the injected hydrogen mass using a trained machine learning model (108) using the calculated pressure drop; c. determining the incremental mass by calculating the difference between the estimated hydrogen mass and the set point mass according to the operating conditions of the engine, and d. Determining the injection mass deviation of the HGI (110) by comparing the incremental mass to a threshold value.

2. The apparatus of claim 1 , wherein the machine learning model (108) is trained using parameters including engine speed, fuel rail pressure before injection, fuel rail pressure drop due to injection, fuel rail temperature, injector energization duration, engine torque, and mass flow through a hydrogen injector pressure regulator (104).

3. The apparatus of claim 1, wherein the determined deviation is compared with an allowable range and based thereon, a correction of the HGI (110) is performed.

4. The apparatus of claim 1, wherein the determined deviation of the HGI (110) is transmitted over a cloud to update the HGI (110) state. 5 . The device of claim 1 , which is at least one selected from the group consisting of a smartphone, a computer, an ECU, and a cloud.

6. A method for determining injection quality deviation in a hydrogen gas injector (HGI) (110), the method comprising the steps of: a. Measuring real-time hydrogen pressure in a hydrogen rail (106) fluidly connected to the HGI (110), wherein the hydrogen pressure is measured before and after injection into a combustion chamber (112) of an engine, characterized in that: b. calculating the pressure drop in the hydrogen rail (106) and estimating the injected hydrogen mass using the trained machine learning model (108) using the calculated pressure drop; c. determining the incremental mass by calculating the difference between the estimated hydrogen mass and the set point mass according to the operating conditions of the engine, and d. Determining the shot quantity deviation of the HGI (110) by comparing the incremental mass with a threshold value.

7. The method of claim 6, wherein the machine learning model 108 is trained using parameters including engine speed, fuel rail pressure before injection, fuel rail pressure drop due to injection, fuel rail temperature, injector energization, engine torque, and mass flow through a hydrogen injector pressure regulator (104).

8. The method of claim 6, comprising: When the determined injection mass deviation is within an allowable range, the operation of the HGI (110) is corrected.

9. The method of claim 6, comprising: transmitting the determined injection mass deviation of the HGI (110) to a cloud; and updating the HGI (110) state.

10. The method of claim 6, performed by a controller (100), wherein the controller (100) is at least one selected from the group consisting of a smartphone, a computer, an ECU, and a cloud.

Images