System monitoring refrigeration charge

• US 5,860,286 A
• Filed: 06/06/1997
• Issued: 01/19/1999
• Est. Priority Date: 06/06/1997
• Status: Expired due to Term

• Current Assignee: Carrier Corporation (Carrier Global Corporation)
• Original Assignee: Carrier Corporation (Carrier Global Corporation)
• Inventors: Tulpule, Sharayu
• Primary Examiner(s): Tanner, Harry B.
• Application Number:
  US08/869,536
• Time in Patent Office: 592 Days
• Field of Search: 62/125, 62/126, 62/127, 62/129, 702/33, 702/34, 706/15, 706/38, 706/41, 706/20, 706/23, 706/25
• US Class Current: 62/129
• CPC Class Codes: F25B 2400/06 Several compression cycles arranged in parallel
• CPC Class Codes: F25B 2700/04 Refrigerant level
• 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
• CPC Class Codes: G05B 23/0254 based on a quantitative model, e.g. mathematical relationships between inputs and outputs; functions: observer, Kalman filter, residual calculation, Neural Networks

Abstract

A refrigerant monitoring system for a heating or cooling system includes a neural network that is used to compute the refrigerant charge in at least one refrigeration circuit of the system. The neural network is trained to learn certain characteristics of the heating or cooling system during a development mode of operation. The thus trained neural network timely computes refrigerant charge during a run time mode of operation. Information as to the computed refrigerant charge being at variance with the nominal amount of refrigerant charge for at least one refrigeration circuit is made available for assessment during the run time mode of operation.

Claims

1. A process for monitoring the refrigerant charge in at least one refrigeration circuit of a heating or cooling system comprising the steps of:

reading values of sensed conditions occurring in the heating or cooling system wherein the values are produced by a plurality of sensors positioned at various locations within the heating or cooling system;

processing the read values of the sensed conditions occurring in the heating or cooling system through a neural network so as to produce a computed refrigerant charge for at least one refrigeration circuit of the heating or cooling system that is based on the read values of sensed conditions;

comparing the computed refrigerant charge with a nominal refrigerant charge for at least one refrigeration circuit of the heating or cooling system; and

transmitting a status message as to the amount of refrigerant charge in at least one refrigeration circuit of the refrigeration circuit in response to said step of comparing the computed refrigerant charge with the nominal refrigerant charge.

2. The process of claim 1 wherein the neural network comprises a first array of nodes wherein each node in the first array contains stored values that have been learned by the neural network for particular amounts of refrigerant charge in at least one refrigeration circuit of the heating or cooling system and wherein said step of processing the read values of the sensed conditions through a neural network comprises the step of:

determining the proximity of the read values of sensed conditions occurring in the heating or cooling system to the stored learned values at each of the nodes in the first array; and

identifying a number of nodes in the first array that are closest to the read values of sensed conditions occurring in the heating or cooling system based upon the determined proximity of the read values of sensed conditions to the stored learned values of the nodes.

3. The process of claim 2 wherein the neural network further comprises an array of interpolation nodes wherein each interpolation node is connected to the nodes in the first array through weighted connections that have been previously learned by the neural network, said process further comprising the step of:

computing values at each interpolation node based upon the values of the weighted connections of each interpolation node to the identified number of nodes in the first array.

4. The process of claim 3 wherein the neural network further comprises at least one output node that is connected to each interpolation node through weighted connections that have been previously learned by the neural network, said process further comprising the step of:

computing an output refrigerant charge value based upon both the values of the weighted connections of the output node to each interpolation node and the computed values of each interpolation node.

5. The process of claim 2 wherein the particular amounts of refrigerant charge in at least one refrigeration circuit include at least two amounts of refrigerant charge wherein one of the two amounts of refrigerant charge is the nominal amount of refrigerant charge in at least one refrigeration circuit of the heating or cooling system and the other amount of refrigerant charge is below the nominal amount of refrigerant charge.

6. The process of claim 2 wherein the particular amounts of refrigerant charge in at least one refrigeration circuit include at least two amounts of refrigerant charge wherein one of the two amounts of refrigerant charge is the nominal amount of refrigerant charge in at least one refrigeration circuit of the heating or cooling system and the second amount lies below the nominal amount of refrigerant charge.

7. The process of claim 1 wherein said step of reading values of sensed conditions occurring in the heating or cooling system comprises the steps of:

reading the value of at least one sensed condition occurring with respect to at least one compressor in at least one refrigeration circuit of the heating or cooling system;

reading the value of at least one sensed condition occurring with respect to a heat exchanger in at least one refrigeration circuit of the heating or cooling system; and

reading the value of at least one sensed temperature condition of the refrigerant downstream of the heat exchanger and upstream of an expansion valve in the refrigeration circuit of the heating or cooling system.

8. The process of claim 7 wherein said step of reading the value of at least one condition occurring with respect to at least one compressor in at least one refrigeration circuit comprises the steps of:

reading the compressor suction temperature at the inlet of the compressor; and

reading the compressor discharge temperature at the outlet of the compressor.

9. The process of claim 7 wherein said step of reading the value of at least one sensed condition occurring with respect to at least one heat exchanger in at least one refrigeration circuit of the heating or cooling system comprises the steps of:

reading the temperature of air before entering the heat exchanger; and

reading the temperature of the air leaving the heat exchanger.

10. The process of claim 7 further comprising the steps of:

reading the value of at least one sensed condition occurring with respect to at least one compressor in at least a second refrigeration circuit of the heating or cooling system; and

reading the value of at least one sensed condition of a refrigerant in the second refrigeration circuit upstream of an expansion valve for the second refrigeration circuit within the heating or cooling system.

11. A process for learning the characteristics of a heating or cooling system so as to predict the refrigerant charge in at least one refrigeration circuit of the heating or cooling system, said process comprising the steps of:

storing a plurality of sets of data for sensed conditions occurring within the heating or cooling system when the system is subjected to various load and ambient conditions for known amounts of refrigerant charge in one or more refrigeration circuits of the system; and

repetitively processing a number of the stored sets of data through a neural network so as to teach the neural network to accurately predict known refrigerant charges for the particular sets of data in at least one refrigeration circuit whereby the neural network may be used thereafter to process sensed conditions occurring within the heating or cooling system for unknown refrigerant charges so as to produce a computed refrigerant charge in at least one refrigeration circuit for the sensed set of conditions.

12. The process of claim 11 wherein the neural network comprises a plurality of nodes arranged in a first array wherein each node has a set of weighted values assigned thereto, and wherein the neural network further comprises a plurality of interpolation nodes in a second array wherein the interpolation nodes in the second array have weighted connections to the nodes in the first array and an output node for computing refrigerant charge, the output node having weighted connections to the interpolation nodes in the second array.

13. The process of claim 12 further comprising the step of:

adjusting the weighted connections between the nodes of the first array and the interpolation nodes in the second array in response to the repetitive processing of the number of stored sets of data; and

adjusting the weighted connections between the interpolation nodes of the second array and the output node in response to the repetitive processing of the number of stored sets of data; and

computing new values of refrigerant charge based upon the adjusted weighted connections whereby the adjusted weighted connections between all nodes eventually produce a computed refrigerant charge that converges to the known value of refrigerant charge for the sets of data being respectively processed through the neural network.

14. The process of claim 11 wherein said step of storing a plurality of sets of data for sensed conditions occurring within the heating or cooling system comprises the steps of:

storing each set of data as a plurality of values representing sensed values generated by sensors within the heating or cooling system; and

storing the known refrigerant charge present in the heating or cooling system when the sensed values were generated by the sensors.

15. The process of claim 14 wherein said step of processing a number of stored sets of data through a neural network so as to teach the neural network to predict the known refrigerant charges comprises the steps of:

(a) reading a set of data and storing the read set of data as a set of coordinate dimensions in multidimensional space;

(b) identifying a plurality of nodes in the first array of nodes that possess sets of weighted values close to the set of coordinate dimensions in multidimensional space;

(c) adjusting the sets of weighted values at the identified nodes in the first array whereby the adjusted sets of values are adjusted as a function of the difference between the set of weighted values at each identified node and the coordinate dimensions of the read set of data.

16. The process of claim 15 wherein the sets of weighted values at the identified nodes in the cluster array are adjusted as a function of the value of a learning rate that is multiplied times the difference between the sets of old weighted values at each identified node and the coordinate dimensions values of the read set of data.

17. The process of claim 15 wherein said "a", "b" and "c" steps are repeated until a predetermined number of sets of data stored in the storage device have been processed and wherein said process further comprises the steps of:

(d) reducing the value of the learning rate; and

(e) inquiring as to whether the value of the learning rate has reached a predetermined value.

18. The process of claim 17 wherein said steps of "a", "b", "c", "d" and "e" are repeated until such time as the learning rate has reached the predetermined value, said process furthermore comprising the step of:

storing the finally adjusted weighted values at all nodes in the first array when the value of the learning rate has reached the predetermined value.

19. The process of claim 12 further comprising the step of:

reading a set of data;

adjusting the defined plurality of connections between the nodes of the first array and the interpolation nodes in the second array in response to the read set of data; and

adjusting the defined plurality of connections between the interpolation nodes of the second array and the output node in response to the read set of data whereby the adjusted connections between all nodes eventually produce computed refrigerant charges that converge to the known values of refrigerant charges for the sets of data being repetitively processed.

20. The process of claim 11 wherein said step of storing a plurality of sets of data for sensed conditions occurring within the heating or cooling system comprises the steps of:

storing each set of data as a plurality of values representing sensed values generated by sensors within the heating or cooling system for a known refrigerant charge; and

storing the known refrigerant charge that was present in at least one refrigeration circuit in the heating or cooling system when the sensors generated the particular set of values in association with the respective set of stored sensed values generated by the sensors whereby the known refrigerant charge can be associated with the respective stored set of sensed values.

21. The process of claim 20 wherein said step of storing each set of data as a plurality of values representing sensed values generated by sensors within the heating or cooling system comprises the steps of:

storing at least one sensed value generated by a sensor mounted at the inlet side of a compressor within the heating or cooling system;

storing at least one sensed value generated by a sensor mounted at the outlet side of the compressor;

storing at least one sensed value generated by a sensor measuring the temperature of air entering a heat exchanger within the heating or cooling system; and

storing at least one sensed value generated by a sensor measuring the temperature of air leaving a heat exchanger within the heating or cooling system.

22. The process of claim 21 wherein said step of storing each set of data as a plurality of values representing sensed values generated by sensors within the heating or cooling system comprises the steps of:

storing at least one sensed value of refrigerant temperature generated by a sensor mounted relative to an expansion valve within the heating or cooling system.

23. A process for monitoring the refrigerant charge present in at least one refrigeration circuit of a heating or cooling system comprising the steps of:

repetitively reading values of certain sensed conditions produced by a plurality of sensors positioned at various locations within the heating or cooling system;

storing each set of read values in a predetermined manner so that each stored value in a set of read values defines a coordinate dimension in multidimensional space;

processing each stored set of values through a neural network whereby a computed refrigerant charge is produced at the output of the neural network for each stored set of read values;

storing each computed refrigerant charge produced at the output of the neural network for each set of values processed through the neural network; and

computing an average of the stored computed refrigerant charges after a predetermined number of computed refrigerant charges have been produced at the output node.

24. The process of claim 23 further comprising the step of:

comparing the computed average of the stored computed refrigerant charges with a known nominal refrigerant charge for at least one refrigeration circuit within the heating or cooling system; and

generating a message when the computed average of the stored refrigerant charges is below the nominal refrigerant charge.

25. The process of claim 23 further comprising the step of:

comparing the computed average of the stored computed refrigerant charges with a known nominal refrigerant charge for at least one refrigeration circuit within the heating or cooling system; and

generating a message when the computed average of the stored refrigerant charges is above the nominal refrigerant charge.

26. The process of claim 23 further comprising the step of:

repeating said steps of repetitively reading values of certain sensed conditions, storing each set of read values, and processing each stored set of read values through a neural network whereby a new computed refrigerant charge is produced for each processed set of read values; and

storing each new computed refrigerant charge for each processed set of values; and

computing an average of the stored new computed refrigerant charges.

27. The process of claim 23 wherein the neural network comprises a plurality of nodes in a first array of nodes wherein each node in the first array contains stored values that have been learned by the neural network and wherein said step of processing the read values of the sensed conditions through a neural network comprises the step of:

determining the proximity of the read and stored values of sensed conditions produced by the plurality of sensors in the heating or cooling system to the stored learned values at each of the nodes in the first array; and

identifying a number of nodes in the first array that are closest to the read and stored values of sensed conditions produced by the plurality of sensors in the heating or cooling system based upon the determined proximity of the read and stored values of sensed conditions to the stored learned values of these nodes.

28. The process of claim 27 wherein the neural network further comprises an array of interpolation nodes wherein each interpolation node is connected to the nodes in the first array through weighted connections that have been previously learned by the neural network, said process further comprising the step of:

computing values at each interpolation node based upon the values of the weighted connections of each interpolation node to the identified number of nodes in the first array.

29. The process of claim 28 wherein the neural network further comprises at least one output node that is connected to each interpolation node through weighted connections that have been previously learned by the neural network, said process further comprising the step of:

computing an output refrigerant charge value based upon the values of the weighted connections to each interpolation node and the computed values of each of the interpolation nodes in the first array.

30. The process of claim 23 wherein the neural network has previously learned neural network values for at least two refrigerant charges in at least one refrigeration circuit of the heating or cooling system wherein one of the two refrigerant charges is the nominal refrigerant charge for at least one refrigeration circuit of the heating or cooling system and the other refrigerant charge lies below the nominal refrigerant charge wherein said step of processing each set of read values through the network comprises:

using the set of read values of sensed conditions occurring in the heating or cooling system to interpolate between learned neural network values so as to produce the computed refrigerant charge.

31. The process of claim 23 wherein the neural network has previously learned neural network values for at least two refrigerant charges in at least one refrigeration circuit of the heating or cooling system wherein one of the two refrigerant charges is the nominal refrigerant charge and the other refrigerant charge lies above the nominal refrigerant charge for at least one refrigeration circuit of the heating or cooling system wherein said step of processing each set of read values through the neural network comprises:

using the set of read values of sensed conditions occurring in the heating or cooling system to interpolate between the learned neural network values so as to produce the computed refrigerant charge.

32. The process of claim 23 wherein the neural network has previously learned neural network values for at least three amounts of refrigerant charge in the refrigeration circuit wherein one of the three refrigerant charges is the nominal refrigerant charge for at least one refrigerant circuit of the heating or cooling system and the other two amounts of refrigerant charge lie to either side of the nominal refrigerant charge wherein said step of processing each set of read values through the neural network comprises the step of:

using the set of read values of sensed conditions occurring in the heating or cooling system to interpolate between the learned neural network values so as to produce the computed refrigerant charge.