Process for refrigerant charge level detection using a neural net

• US 7,111,469 B2
• Filed: 09/22/2004
• Issued: 09/26/2006
• Est. Priority Date: 09/22/2004
• Status: Active Grant

• Current Assignee: Northrop Grumman Systems Corporation (Northrop Grumman Corporation)
• Original Assignee: Northrop Grumman Corporation
• Inventors: Horan, Christopher J., Stasi, William J.
• Primary Examiner(s): NORMAN, MARC E
• Application Number:
  US10/947,483
• Publication Number:
  US 20060059925A1
• Time in Patent Office: 734 Days
• Field of Search: 621/26, 621/29, 621/49, 622/08, 622/09
• US Class Current: 62/129
• CPC Class Codes: B60H 1/00585 Means for monitoring, testing or servicing the air-conditioning
• CPC Class Codes: B64D 13/06 the air being conditioned pressurising B64D13/02
• CPC Class Codes: F25B 2400/13 Economisers
• CPC Class Codes: F25B 2500/19 Calculation of parameters
• CPC Class Codes: F25B 2700/04 Refrigerant level
• CPC Class Codes: F25B 2700/1931 Discharge pressures
• CPC Class Codes: F25B 2700/21161 of the fluid heated by the condenser
• CPC Class Codes: F25B 2700/21163 of the refrigerant at the outlet of the condenser
• CPC Class Codes: F25B 2700/21172 at the inlet
• CPC Class Codes: F25B 2700/21173 at the outlet
• CPC Class Codes: F25B 45/00 Arrangements for charging or discharging refrigerant
• CPC Class Codes: F25B 49/005 of safety devices F25B49/02 and F25B49/04 take precedence

Abstract

The invention is a process for determining the charge level of a vapor cycle environmental control system, having a condenser, evaporator, and an expansion valve, comprising the steps of providing a neural network having four input neurons, two hidden neurons and three output neurons; determining the number of degrees below the saturation temperature of the liquid refrigerant: exiting the condenser and providing this measurement to the first input neuron; sensing the condenser sink temperature and providing the measurement to the second input neuron; sensing either the refrigerant outlet temperature from the condenser or the evaporator exhaust air temperature and providing the measurement to the third input neuron, sensing the evaporator inlet temperature and providing the measurement to the fourth input neuron; and using the trained neural network to monitor the charge level in the system.

Claims

1. A process for determining the charge level of a vapor cycle environmental control system, having a condenser, evaporator, compressor, and an expansion valve comprising the steps of:

providing a neural network having four input neurons, two hidden neurons and three output neurons;

determining the number of degrees below the saturation temperature of the liquid refrigerant exiting the condenser and providing this measurement to the first input neuron;

sensing the condenser sink temperature and providing the measurement to the second input neuron;

sensing either the refrigerant outlet temperature from the condenser or the evaporator exhaust air temperature and providing the measurement to the third input neuron;

sensing the evaporator air inlet temperature and providing the measurement to the fourth input neuron;

training the neural network by providing under, over and normal refrigerant charge levels to the system and operating the system under varying operating conditions such that weighting factors for the neural network are determined; and

using the trained neural network to monitor the charge level in the system.

2. The process as set forth in claim 1 wherein the step of training the neural network by providing under, over and normal refrigerant charge levels to the system and operating the system under varying operating conditions such that weighting factors for the neural network are determined includes the step of applying a log-sig function to each output from each hidden neuron.

3. The process as set forth in claim 1 wherein the step of training the neural network by providing under, over and normal refrigerant charge levels to the system and operating the system under varying operating conditions such that weighting factors for the neural network are determined includes the step of taking an N-point moving average, rounding the result, and requiring N of M of these rounded values to agree before reporting the charge state estimate are defined.

4. A process for determining the charge level of a vapor cycle environmental control system, having a condenser, evaporator, compressor, and an expansion valve comprising the steps:

determining the number of degrees below the saturation temperature of the liquid refrigerant exiting the condenser;

sensing the condenser sink temperature;

sensing either the refrigerant outlet temperature from the condenser or the evaporator exhaust air, and sensing the evaporator air inlet and providing these inputs to a neural network;

training the neural network by providing under, over and normal refrigerant charge levels to the system and operating the system under varying operating conditions such that weighting factors for the neural network are determined; and

using the trained neural network to monitor the charge level in the system.

5. The process as set forth in claim 4 wherein the step of providing inputs to a neural network comprises providing a neural network which includes four input neurons, two hidden neurons and three output neurons.

6. The process as set forth in claim 5 wherein the step of training the neural network by providing under, over and normal refrigerant charge levels to the system and operating the system under varying operating conditions such that weighting factors for the neural network are determined includes the step of applying a log-sig function to each output from each hidden neuron.

7. The process as set forth in claim 6 wherein the step of training the neural network by providing under, over and normal refrigerant charge levels to the system and operating the system under varying operating conditions such that weighting factors for the neural network are determined includes the step of taking an N-point moving average, rounding the result, and requiring N of M of these rounded values to agree before reporting the charge state estimate are defined.