Adaptive temperature control chiller system control method and chiller system

Abstract

The application discloses a kind of self-adapting temperature control chiller system control method and chiller system, method includes the following steps: S1, the operating parameter of real-time acquisition chiller, at least including current exhaust temperature, exhaust temperature variation rate, lubricating oil temperature, compressor frequency and ambient temperature;S2, according to current operating mode, call adaptive threshold prediction model, obtain exhaust temperature dynamic threshold, the exhaust temperature dynamic threshold is obtained according to the compressor frequency, ambient temperature and lubricating oil temperature dynamic adjustment calculation coefficient;S3, based on historical operating parameter, predict the predicted exhaust temperature of future time;S4, according to the comparison result of current exhaust temperature, predicted exhaust temperature and exhaust temperature too low threshold, self-adapting adjustment electronic expansion valve opening degree and / or compressor frequency.Can solve the problem of chiller exhaust temperature too low under low temperature environment, leading to compressor oil shortage wear and efficiency decline.

Description

Technical Field

[0001] This invention relates to the field of temperature-controlled chiller technology, specifically to an adaptive temperature-controlled chiller system control method and chiller system. Background Technology

[0002] In low-temperature operating conditions, excessively low compressor discharge temperature is a key hidden danger leading to decreased unit reliability and energy efficiency degradation. When the discharge temperature falls below the safe threshold, the lubricating oil inside the compressor is diluted by a large amount of refrigerant, causing a sharp drop in viscosity and preventing the formation of an effective oil film on the friction surfaces. Simultaneously, the lubricating oil dissolved in the refrigerant migrates with the discharge into the system piping, resulting in a decrease in the compressor's internal oil level and deterioration of lubrication. This can lead to serious mechanical failures such as bearing wear, rotor seizure, and even motor burnout. More seriously, excessively low discharge temperature also significantly reduces refrigerant circulation efficiency, resulting in decreased heating / cooling capacity and increased energy consumption.

[0003] However, traditional temperature control methods often employ fixed thresholds for passive protection. For example, Chinese patent CN118517778A only triggers alarms and adjustments after the exhaust temperature has fallen below a critical value. This "post-event response" mode has inherent flaws. Due to the large hysteresis and strong coupling nonlinear thermodynamic characteristics of refrigeration systems, the response cycle from temperature deviation to triggering protection, and then to the adjustment command taking effect and changing the system state, often requires tens of seconds or even minutes. During this period, the compressor is already in a dangerous state of insufficient lubrication for an extended period, and irreversible damage to lubricating oil and mechanical wear accumulate. By the time the system completes adjustment, the optimal intervention opportunity has already been missed. Therefore, it is urgent to break through the limitations of traditional passive control and establish a predictive control mechanism that can anticipate temperature change trends and proactively intervene before danger occurs, achieving a technological leap from "locking the stable door after the horse has bolted" to "prevention is better than cure." Summary of the Invention

[0004] This invention proposes an adaptive temperature control chiller system control method and chiller system, which can solve the problems of compressor oil shortage, wear and efficiency reduction caused by excessively low chiller exhaust temperature in low temperature environments.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: an adaptive temperature control chiller system control method, comprising the following steps: S1, real-time acquisition of the chiller's operating parameters, including at least the current exhaust temperature, exhaust temperature change rate, lubricating oil temperature, compressor frequency, and ambient temperature; S2, based on the current operating mode, call the adaptive threshold prediction model to obtain the dynamic threshold of exhaust temperature. The dynamic threshold of exhaust temperature is obtained by dynamically adjusting the calculation coefficients based on the compressor frequency, ambient temperature and lubricating oil temperature. S3, based on historical operating parameters, predicts the exhaust temperature at future moments; S4, based on the comparison result of the current exhaust temperature, the predicted exhaust temperature and the exhaust temperature too low threshold, adaptively adjust the opening of the electronic expansion valve and / or the compressor frequency.

[0006] In step S1, the operating parameters may include the compressor's real-time operating frequency, exhaust temperature, previous cycle exhaust temperature, exhaust pressure, exhaust saturation temperature, lubricating oil temperature, ambient temperature, water outlet temperature or inlet temperature of the water heat exchanger, and the exhaust temperature change rate.

[0007] Preferably, in step S2, the low exhaust temperature threshold is determined by the superposition of multiple dynamic correction terms. The dynamic correction terms include at least: a load base term, which is related to the real-time frequency of the compressor; an oil temperature correction term, which is related to the real-time temperature of the lubricating oil to compensate for the risk of lubricating oil flowability; an inlet / outlet water temperature term, which is related to the outlet or inlet water temperature of the water heat exchanger; an ambient temperature term, which is related to the ambient temperature; and a rate of change correction term, which is related to the trend of exhaust temperature change. The dynamic threshold of exhaust temperature is expressed as a polynomial composed of the load base term, the oil temperature correction term, the inlet / outlet water temperature term, the ambient temperature term, and the rate of change correction term.

[0008] Preferably, in step S2, the load base item, inlet / outlet water temperature item, and ambient temperature item are each configured with a dynamic adjustment coefficient to achieve the following: the load base item increases as the compressor load increases; the ambient temperature item increases as the ambient temperature decreases in heating mode and increases as the ambient temperature increases in cooling mode; the inlet / outlet water temperature item increases as the outlet water temperature increases in heating mode and increases as the inlet water temperature decreases in cooling mode.

[0009] Preferably, in step S2, the oil temperature correction term and the rate of change correction term are dynamically determined in terms of correction direction and magnitude based on real-time operating conditions; in heating mode, the load base term and the inlet and outlet water temperature term have a positive correction effect on the threshold; the ambient temperature term has a negative correction effect on the threshold; in cooling mode, the load base term and the ambient temperature term have a positive correction effect on the threshold; the inlet and outlet water temperature term has a negative correction effect on the threshold; this means that in heating mode, the inlet and outlet water temperature term is an addition term and the ambient temperature term is a subtraction term; in cooling mode, the ambient temperature term is an addition term and the inlet and outlet water temperature term is a subtraction term.

[0010] Preferably, in step S2, the correction magnitude of the oil temperature correction term is negatively correlated with the lubricating oil temperature. When the lubricating oil temperature is lower than the normal operating range, the correction magnitude of the oil temperature correction term increases as the temperature decreases.

[0011] Preferably, in step S2, the rate of change correction term is configured with a trend sensitivity coefficient; when the exhaust temperature shows a downward trend, the trend sensitivity coefficient increases as the absolute value of the rate of decrease increases; when the exhaust temperature shows an upward trend or is stable, the trend sensitivity coefficient maintains a baseline value or decreases.

[0012] Preferably, in step S3, the predicted exhaust temperature for the future moment is achieved through a time-series prediction model of a long short-term memory network. The historical operating data includes time-series data of exhaust temperature, ambient temperature, and compressor frequency for the past N collection cycles, and the future moment is the predicted moment 5-10 seconds later.

[0013] Preferably, step S4 includes: If the exhaust temperature or predicted exhaust temperature is lower than the warning threshold and remains below the first preset duration, the opening of the electronic expansion valve is locked and can only be reduced, and the compressor frequency is reduced. If the exhaust temperature or predicted exhaust temperature is below the intervention threshold and remains below the second preset duration, the opening of the electronic expansion valve is dynamically adjusted according to the temperature change trend. If both the exhaust temperature and the predicted exhaust temperature are higher than the recovery threshold and remain above the third preset duration, the low exhaust temperature control will be exited, and the system will switch to dual-target optimization control of supplementary air temperature and energy consumption.

[0014] The present invention also adopts the following technical solution: an adaptive temperature control chiller system, employing the above-mentioned adaptive temperature control chiller system control method, comprising: a four-way reversing valve with port C connected to a water heat exchanger, port E connected to a finned heat exchanger, and ports D and S connected to a compressor exhaust port and intake port; a flash evaporator, with its inlet and outlet connected to the water heat exchanger and finned heat exchanger respectively via heating and cooling electronic expansion valves, and an air supply port connected to the compressor air supply port via an air supply solenoid valve; and a control module connected to a sensor group that collects signals from the compressor operating status and environmental parameters, and outputting control commands to the chiller system.

[0015] Preferably, the sensor group includes: an exhaust temperature sensor, an exhaust pressure sensor, and a lubricating oil temperature sensor installed on the compressor; an inlet water temperature sensor and an outlet water temperature sensor installed on the water heat exchanger; and an ambient temperature sensor installed on the finned heat exchanger. The control module outputs control commands to the compressor, the heating electronic expansion valve, the cooling electronic expansion valve, and the gas injection solenoid valve.

[0016] The beneficial effects of this invention are: 1) Significantly improved control accuracy: Through multi-parameter fusion and AI predictive control, exhaust temperature fluctuations are controlled within ±1℃, and lubricating oil temperature is stabilized within a safe range, effectively solving the problem of excessive temperature fluctuations in traditional methods; 2) Enhanced Adaptability: The dynamic coefficient adjustment mechanism replaces the fixed coefficient, enabling the system to automatically adapt to different operating conditions such as variable load and extreme temperature, achieving optimized operation under all operating conditions without manual intervention; 3) Coordinated optimization of energy consumption and reliability: Energy consumption weight is introduced into the temperature control process, and energy saving of 5%-10% is achieved through dual-objective optimization control; at the same time, the temperature change trend is predicted in advance to avoid lubricating oil loss and extend the unit life by more than 30%; 4) Improved response speed: The control decision time is adaptively adjusted based on the rate of temperature change, which shortens the response time under extreme conditions and achieves rapid protection; the intervention-level adjustment cycle is shortened, which improves the dynamic response performance of the system. Attached Figure Description

[0017] Figure 1 This is a flowchart of an adaptive temperature control chiller system control method according to Embodiment 1 of the present invention.

[0018] Figure 2 This is a structural block diagram of an adaptive temperature control chiller system according to the present invention.

[0019] Reference numerals: 1. Compressor; 2. Four-way reversing valve; 3. Finned heat exchanger; 4. Water heat exchanger; 5. Flash evaporator; 6. Heating electronic expansion valve; 7. Cooling electronic expansion valve; 8. Gas injection solenoid valve; 9. Exhaust pressure sensor; 10. Exhaust temperature sensor; 11. Suction pressure sensor; 12. Suction temperature sensor; 13. Inlet water temperature sensor; 14. Outlet water temperature sensor; 15. Ambient temperature sensor; 16. Lubricating oil temperature sensor; 17. Control module; 18. Liquid storage tank; 19. Gas-liquid separator. Detailed Implementation

[0020] Example 1

[0021] This embodiment provides an adaptive temperature control chiller system control method and chiller system, which is applicable to low-temperature EVI heat pump chiller units and various industrial or civil temperature control scenarios, and can achieve precise control of exhaust temperature, optimization of unit performance and extension of service life.

[0022] Reference Figure 1 The control method in this embodiment includes the following steps.

[0023] Step S1: Collect the operating parameters of the chiller in real time.

[0024] After the unit is started, the control module 17 automatically identifies the operating mode (cooling / heating) and initiates a multi-parameter real-time acquisition process. The operating parameters include at least the current exhaust temperature, exhaust temperature change rate, lubricating oil temperature, compressor frequency, and ambient temperature.

[0025] Specifically, the control module 17 acquires data from each sensor every preset acquisition cycle, in this embodiment T1=3s.

[0026] Compressor operating status parameters: The current exhaust temperature Td is collected by the exhaust temperature sensor 10 located on compressor 1, the exhaust pressure is collected by the exhaust pressure sensor 9 and converted into the exhaust saturation temperature Tpd, the lubricating oil temperature Toil is collected by the lubricating oil temperature sensor 16, and the current operating frequency Fx is read directly from the compressor frequency converter; Heat exchanger status parameters: The inlet water temperature Tin and the outlet water temperature Tout are collected by the inlet water temperature sensor 13 and the outlet water temperature sensor 14 installed in the water heat exchanger 4, respectively. Environmental status parameters: Ambient temperature Tao is collected by ambient temperature sensor 15 located in finned heat exchanger 3; Historical data recording: Control module 17 synchronously stores the exhaust temperature Td of the previous cycle. n-1 , is used to calculate the rate of change of exhaust temperature dTd / dt.

[0027] The exhaust temperature change rate dTd / dt is obtained through differential calculation, i.e. (Td n - Td n-1 ) / T1, where Td n This is the exhaust temperature for this cycle.

[0028] All sensor signals are converted from analog to digital and then fed into the data acquisition unit of control module 17, providing a real-time data foundation for subsequent dynamic threshold calculation and AI prediction. After acquisition is complete, proceed to step S2.

[0029] Step S2: Based on the current operating mode, call the adaptive threshold prediction model to obtain the dynamic threshold of exhaust temperature.

[0030] Based on the operating mode (cooling / heating) identified in step S1, control module 17 calls the corresponding adaptive threshold prediction model to calculate the dynamic threshold of exhaust temperature. In this embodiment, the low exhaust temperature value Tdoh is calculated in heating mode, and the low exhaust temperature value Tdoc is calculated in cooling mode. The core innovation of the dynamic threshold lies in using an adaptive coefficient instead of a traditional fixed coefficient, optimizing in real time based on compressor frequency, ambient temperature, lubricating oil temperature, and temperature change rate to achieve precise protection under all operating conditions.

[0031] The detailed structure of the dynamic correction term is explained below.

[0032] The dynamic threshold of exhaust temperature is determined by the superposition of five dynamic correction terms. The physical meaning and calculation logic of each correction term are as follows.

[0033] First, the load base item.

[0034] The load baseline item reflects the benchmark impact of the compressor's operating load on the exhaust temperature, with the compressor's real-time frequency Fx as the input variable. The control module 17 establishes a frequency-load mapping table based on unit operating experience data, converting Fx into a relative load rate. Higher loads result in greater compressor work, higher exhaust temperature benchmarks, and a corresponding increase in the required protection threshold. In this embodiment, the load baseline item is represented as Ah(Fx, Tao) × Fx, where Ah is the load dynamic coefficient. Its value is not only related to Fx but also influenced by ambient temperature—compressor efficiency decreases at low temperatures, increasing the actual load at the same frequency, requiring Ah to increase accordingly to compensate for calculation deviations.

[0035] Second, oil temperature correction item.

[0036] The oil temperature correction term is associated with the real-time lubricating oil temperature (Toil) and is used to compensate for lubricating oil flow risks. Lubricating oil temperature directly affects the internal lubrication effect of the compressor: when the temperature is too low, the lubricating oil viscosity increases, flowability decreases, friction loss increases, and the exhaust temperature is prone to abnormal rise, but lubrication reliability actually decreases; when the temperature is too high, the lubricating oil viscosity decreases, oil film strength is insufficient, and the risk of wear increases. In this embodiment, the oil temperature correction term is represented as Bh(Toil), where Bh is the oil temperature correction coefficient.

[0037] The determination of Bh follows these rules: When Toil is within the normal operating range of 30-60℃, Bh maintains a baseline value; when Toil is below 30℃, Bh increases as the temperature decreases, with the rate of increase positively correlated with the degree of temperature deviation, to trigger protection in advance and prevent low-temperature lubrication failure; when Toil is above 60℃, Bh decreases as the temperature increases, with the rate of decrease positively correlated with the degree of temperature deviation, to avoid excessive protection at high temperatures leading to increased energy consumption. The relationship between Bh and Toil is non-linear and negatively correlated; the specific mapping relationship is calibrated through unit bench tests and stored in the lookup table of control module 17.

[0038] Third, the inlet and outlet water temperatures.

[0039] The inlet and outlet water temperatures are related to the water temperature of the water heat exchanger and are used to compensate for the impact of water-side heat exchange on the exhaust temperature. The control module 17 automatically switches the input variables according to the operating mode: the outlet water temperature Tout is used in heating mode, and the inlet water temperature Tin is used in cooling mode.

[0040] In heating mode, the outlet water temperature Tout directly reflects the user's heating demand. A higher Tout indicates a greater heat load demand, requiring corresponding adjustments to the system circulation flow rate or heat exchange temperature difference, and consequently changing the exhaust temperature baseline. In this embodiment, the heating inlet and outlet water temperatures are represented as Ch(Tout) × Tout, where Ch is the outlet water temperature dynamic coefficient. The value of Ch is positively correlated with Tout—the higher Tout, the larger Ch, and the greater the positive correction amplitude of the threshold, ensuring a safety margin for exhaust temperature under high supply water temperature conditions.

[0041] In cooling mode, the inlet water temperature Tin reflects the initial state of the cooling water. A higher Tin results in a greater heat exchange load on the evaporator, increased compressor suction pressure, and a higher exhaust temperature. In this embodiment, the cooling inlet and outlet water temperatures are represented as Dc(Tin)×Tin, where Dc is the dynamic coefficient of the inlet water temperature. The value of Dc is negatively correlated with Tin—the higher the Tin, the smaller Dc (the larger the absolute value). Since this term provides a negative correction in the cooling threshold formula, the actual effect is a lower threshold, preventing excessive protection at high inlet water temperatures that could lead to insufficient cooling capacity.

[0042] Fourth, ambient temperature.

[0043] The ambient temperature term is associated with the ambient temperature Tao and is used to compensate for the impact of the external environment on the system cycle. The ambient temperature directly affects the heat exchange efficiency of the finned heat exchanger 3, which in turn affects the compressor's intake and exhaust temperatures.

[0044] In heating mode, the lower the ambient temperature Tao, the more difficult it is for the finned heat exchanger to absorb heat from the air, resulting in a decrease in evaporation temperature, an increase in compressor pressure ratio, and a risk of excessively low exhaust temperature leading to lubrication problems. In this embodiment, the heating ambient temperature term is represented as -Dh(Tao)×Tao, where Dh is the ambient temperature dynamic coefficient. The value of Dh is negatively correlated with Tao—the lower Tao, the larger Dh. Since this term is a negative correction (-Dh×Tao, which is actually an additive term when Tao is negative), the threshold increase is greater, effectively compensating for the risk of exhaust temperature drop in low-temperature environments.

[0045] In cooling mode, the higher the ambient temperature Tao, the more difficult it is for the finned heat exchanger to dissipate heat to the air, leading to an increase in condensation temperature, increased compressor discharge pressure, and a tendency for the discharge temperature to become excessively high. In this embodiment, the ambient temperature term is represented as Cc(Tao)×Tao, where Cc is the dynamic coefficient of ambient temperature. The value of Cc is positively correlated with Tao—the higher Tao, the larger Cc, and the greater the positive correction amplitude of the threshold, thus triggering protection earlier to prevent compressor overload caused by high temperature and high pressure.

[0046] Fifth, the rate of change correction term.

[0047] The rate of change correction term is correlated with the exhaust temperature change trend dTd / dt, and is used to achieve predictive dynamic compensation. This correction term enables the threshold to have a "trend-sensing" capability, adjusting the protection boundary in advance when the temperature changes rapidly, and avoiding protection lag caused by passive response.

[0048] In this embodiment, the rate of change correction term is represented as Eh(dTd / dt), where Eh is the trend sensitivity coefficient. The configuration rules for Eh are as follows: when dTd / dt < 0 (exhaust temperature decreases), Eh increases with the absolute value of the rate of decrease, increasing the threshold and preventing the risk of a sudden temperature drop in advance; when dTd / dt > 0 (exhaust temperature increases), Eh maintains the baseline value or decreases with the rate of increase, avoiding excessive protection during the temperature recovery process that could affect system efficiency; when dTd / dt ≈ 0 (temperature is stable), Eh takes the baseline value and is not further corrected.

[0049] The mapping relationship between Eh and dTd / dt is optimized in real time through a dynamic programming algorithm. The control module 17 fits the current trend curve based on the historical temperature change data of the past 10 cycles, predicts the temperature trajectory in the next 3-5 seconds, and dynamically adjusts the sensitivity of Eh.

[0050] Based on the above five dynamic correction terms, the control module 17 calculates the dynamic threshold of exhaust temperature for heating mode and cooling mode according to the following logic.

[0051] The value of Tdoh, which is too low for heating exhaust temperature, can be expressed as Ah(Fx,Tao)×Fx + Bh(Toil) + Ch(Tout)×Tout - Dh(Tao)×Tao + Eh(dTd / dt).

[0052] Wherein, Ah(Fx,Tao) is the load dynamic coefficient, which increases under high load and low temperature environment; Bh(Toil) is the oil temperature correction coefficient, the lower the Toil, the larger Bh; Ch(Tout) is the outlet water temperature dynamic coefficient, the higher the Tout, the larger Ch; Dh(Tao) is the ambient temperature dynamic coefficient, the lower the Tao, the larger Dh; Eh(dTd / dt) is the trend sensitivity coefficient, the greater the temperature drop rate, the larger Eh.

[0053] In heating mode, the load base item, oil temperature correction item, and inlet / outlet water temperature item are positively corrected, while the ambient temperature item is negatively corrected. However, since Tao is often negative under heating conditions (temperature below zero degrees Celsius), -Dh×Tao actually makes a positive contribution.

[0054] The value of excessively low refrigeration exhaust temperature Tdoc can be expressed as Ac(Fx,Tin)×Fx + Bc(Toil) + Cc(Tao)×Tao - Dc(Tin)×Tin + Ec(dTd / dt).

[0055] Where Ac(Fx,Tin) is the load dynamic coefficient, which increases under high load; Bc(Toil) is the oil temperature correction coefficient, consistent with the Bh logic; Cc(Tao) is the ambient temperature dynamic coefficient, the higher Tao, the larger Cc; Dc(Tin) is the inlet water temperature dynamic coefficient, the higher Tin, the larger Dc; Ec(dTd / dt) is the trend sensitivity coefficient, consistent with the Eh logic.

[0056] In cooling mode, the load base item, oil temperature correction item, and ambient temperature item are positively corrected, while the inlet and outlet water temperature item is negatively corrected, complementing the ambient temperature item and jointly balancing the thermodynamic effects on the water and air sides.

[0057] The real-time optimization mechanism for dynamic coefficients will be explained in detail below.

[0058] The control module 17 has a built-in AI optimization engine that performs real-time online optimization of the aforementioned dynamic coefficients. The optimization goal is to maximize the unit's COP while ensuring that the lubricating oil temperature is within its optimal operating range, provided that the exhaust temperature safety constraint (Td≥Tmin) is met.

[0059] The input features for coefficient optimization include current Fx, Tao, Toil, Tout / Tin, dTd / dt, historical temperature trajectory over 10 cycles, cumulative unit runtime, and historical fault records.

[0060] The output of coefficient optimization is the correction increments ΔA, ΔB, ΔC, ΔD, and ΔE for each dynamic coefficient, which are optimized within the constraint boundaries using a gradient descent algorithm. The optimization cycle is synchronized with the data acquisition cycle (3 seconds) to ensure that the coefficients are always adapted to the latest operating conditions.

[0061] Smooth mode switching: When the unit switches between cooling and heating modes via the four-way reversing valve 2, the control module 17 detects the reversing signal and employs a coefficient gradual change strategy within 5 cycles before and after the mode switch to avoid control oscillations caused by threshold jumps. The gradual temperature Td_new can be expressed as α×Td_old + (1-α)×Td_target, where α decreases from 1 to 0.

[0062] After completing the calculation of the dynamic threshold of exhaust temperature, the control module 17 uses Tdoh or Tdoc as the judgment criterion for subsequent steps S3 and S4, and proceeds to step S3.

[0063] Step S3: Based on historical operating parameters, predict the exhaust temperature for future moments.

[0064] The control module 17 calls a pre-trained Long Short-Term Memory (LSTM) network time-series prediction model to predict the exhaust temperature at future times based on historical operating parameters. The historical operating data includes exhaust temperature Td, ambient temperature Tao, and compressor frequency Fx time-series data for the past N acquisition cycles. In this embodiment, N is 10, corresponding to a 30-second historical time window.

[0065] The LSTM model comprises an input layer, a hidden layer, and an output layer. The input layer receives a normalized three-dimensional temporal feature vector [Td, Tao, Fx]; the hidden layer uses a bidirectional LSTM structure with 64 memory units to capture the forward trend and backward dependency of the temporal data; the output layer outputs the predicted exhaust temperature Td_pred 8 seconds later.

[0066] Before model inference, control module 17 performs sliding window processing on the input data, removing abnormal jump points and performing smoothing filtering. After prediction, Td_pred and the dynamic threshold calculated in step S2 are input together into step S4 for hierarchical control determination.

[0067] Step S4: Based on the comparison results of the current exhaust temperature, the predicted exhaust temperature and the exhaust temperature too low threshold, adaptively adjust the opening of the electronic expansion valve and / or the compressor frequency.

[0068] Based on the dynamic threshold of exhaust temperature obtained in step S2 and the predicted exhaust temperature Td_pred obtained in step S3, control module 17 establishes a dual-dimensional judgment system of measured and predicted values, and adopts a hierarchical response mechanism to achieve adaptive adjustment. The hierarchical response mechanism includes three levels of control: early warning, intervention, and recovery. Each level of control is triggered based on the degree and duration of deviation between the current exhaust temperature Td, the predicted exhaust temperature Td_pred, and the dynamic threshold, and can adaptively adjust the judgment duration according to the rate of change of exhaust temperature.

[0069] Control module 17 sets three threshold boundaries to form a complete control decision space. The warning threshold Td_warn is the dynamic threshold minus the first offset a; in this embodiment, a = 2°C, corresponding to a state approaching the risk zone. The intervention threshold Td_act is the dynamic threshold minus the second offset b; in this embodiment, b = 5°C, corresponding to a state entering the risk zone. The recovery threshold Td_rec is the same as the warning threshold, but the judgment logic is that if it exceeds this threshold, it corresponds to a state leaving the risk zone.

[0070] Simultaneously, safety constraint boundaries based on predicted exhaust temperature are set. The predicted warning boundary is the exhaust saturation temperature Tpd plus a first safety margin d, which in this embodiment is d=5℃. The predicted intervention boundary is the exhaust saturation temperature Tpd plus a second safety margin c, which in this embodiment is c=3℃.

[0071] The dynamic duration adjustment mechanism is explained below.

[0072] The duration of each control level is adaptively adjusted based on the exhaust temperature change rate dTd / dt. The first preset duration t1 is the warning level judgment duration, with a base value of 8 seconds; when the temperature drop rate exceeds 0.5℃ per second, it is shortened to 5 seconds to achieve rapid warning; when the temperature change is stable, it maintains the base value; when the temperature rises rapidly, it is extended to 10 seconds to avoid false judgment. The second preset duration t2 is the intervention level judgment duration, with a base value of 10 seconds, shortened to 6 seconds only when the temperature drop rate exceeds 1.0℃ per second. The third preset duration t3 is the recovery level judgment duration, fixed at 10 seconds, ensuring that the system fully recovers from the risk state before exiting protection.

[0073] The following section explains the early warning level control.

[0074] The triggering condition is that either of the following is met: the current exhaust temperature is lower than the warning threshold, or the predicted exhaust temperature is lower than the predicted warning boundary, and this condition continues for a first preset duration t1.

[0075] Control actions include: locking the electronic expansion valve opening; locking the heating electronic expansion valve in heating mode and the cooling electronic expansion valve in cooling mode; the locking rule is that only reduction is allowed, and increase is prohibited; the current opening is used as the upper limit, and subsequent adjustments can only be made downwards. Compressor frequency optimization aims to reduce the compressor frequency, but introduces energy consumption weight constraints; if the current energy efficiency ratio is higher than the target value of 90%, moderate frequency reduction is allowed, with a reduction not exceeding 10% of the rated frequency; if the current energy efficiency ratio is lower than the target value of 80%, the current frequency is maintained first, and temperature risks are mitigated by adjusting the gas supply. Status marking: control module 17 marks a low-load warning state and records the operating parameters at the trigger time for subsequent optimization and fault tracing.

[0076] The following section elaborates on intervention-level control.

[0077] The triggering condition is that either the current exhaust temperature is lower than the intervention threshold, or the predicted exhaust temperature is lower than the predicted intervention boundary, and this condition continues for a second preset duration t2.

[0078] Temperature trend is determined by comparing the exhaust temperature of the current cycle with that of the previous cycle. When the temperature rises, it indicates that the early warning control has taken effect, and the current electronic expansion valve opening is maintained unchanged while continuously monitoring the output of the predictive model. If the predicted exhaust temperature shows an upward trend for several consecutive cycles, the system can be switched to the early warning state ahead of schedule. When the temperature continues to fall, it indicates that the current control strength is insufficient, and a periodic adjustment program is initiated with an adjustment cycle of 30 seconds, performing an adjustment of the electronic expansion valve opening every 30 seconds.

[0079] The adjustment range is determined by considering the following factors: the degree of deviation between the current temperature and the dynamic threshold (the greater the deviation, the stronger the compensation requirement); the inertial trend of temperature change (the faster the temperature drops, the stronger the compensation requirement); and lubricating oil temperature compensation (the lower the oil temperature, the larger the adjustment range). The overall adjustment amount is output after amplitude limiting, and a single adjustment does not exceed ±5% of the full opening to prevent overshoot. Simultaneously, the PWM duty cycle of the gas injection solenoid valve is adjusted to optimize the gas injection amount of the flash evaporator. The direction of gas injection adjustment is opposite to the direction of electronic expansion valve opening adjustment to compensate for refrigerant flow loss.

[0080] The recovery level control will be explained below.

[0081] The triggering conditions must be met simultaneously: the current exhaust temperature is higher than the recovery threshold, the predicted exhaust temperature is higher than the predicted warning boundary, and the condition continues for a third preset duration of 10 seconds.

[0082] The control actions include exiting the low exhaust temperature protection state, releasing the electronic expansion valve opening lock, and switching the control target to dual-target optimization control of replenishment air temperature and energy consumption.

[0083] The target gas injection temperature is automatically calculated based on the current operating conditions. In heating mode, it is determined by a combination of the outlet water temperature and the ambient temperature, following the principle that the higher the outlet water temperature and the lower the ambient temperature, the higher the target gas injection temperature. In cooling mode, it is determined by a combination of the inlet water temperature and the ambient temperature. The electronic expansion valve opening prioritizes compensating for gas injection temperature deviations. When the absolute value of the deviation is greater than 2℃, gas injection temperature control takes precedence, with an adjustment step of ±3% of the full opening. When the absolute value of the deviation does not exceed 2℃, energy consumption optimization is initiated.

[0084] The energy consumption optimization objective, under the premise of achieving the target gas injection temperature, is to jointly optimize compressor efficiency and minimize the throttling loss of the electronic expansion valve. Compressor efficiency is evaluated through real-time calculation of the energy efficiency ratio. Heating capacity is calculated from the inlet and outlet water temperature difference and flow rate of the water heat exchanger, and input power is calculated from the compressor voltage, current, and power factor. Fine-tuning of the electronic expansion valve opening employs a hill-climbing algorithm, making tentative adjustments in increments of ±1%, continuously fine-tuning along the direction of increasing energy efficiency ratio, with an optimization cycle performed every 60 seconds.

[0085] The dynamic allocation of dual-objective weights is achieved through a comprehensive cost function, with a default weight of 0.6 for gas supply temperature and 0.4 for energy consumption. When the gas supply temperature deviation exceeds 5°C, gas supply temperature is forced to take priority; when the energy efficiency ratio is lower than the energy efficiency limit, energy consumption is forced to take priority. By minimizing the comprehensive cost function, the control module 17 dynamically determines the optimal opening of the electronic expansion valve, achieving coordinated optimization of temperature control and energy-saving operation.

[0086] After completing the adjustment in step S4, the control module 17 returns to step S1 to continue the parameter acquisition and closed-loop control for the next cycle, forming a complete adaptive control loop of acquisition, calculation, prediction, and adjustment. Example 2

[0087] This embodiment provides an adaptive temperature control chiller system that uses the control method described in Embodiment 1. Through multi-sensor fusion data acquisition and intelligent decision-making by the AI ​​control module, it achieves precise control of exhaust temperature and optimization of unit performance.

[0088] Reference Figure 2 The chiller system in this embodiment includes a refrigeration cycle loop, a sensor group, and a control module 17.

[0089] The refrigeration cycle circuit includes a compressor 1, a four-way reversing valve 2, a finned heat exchanger 3, a water heat exchanger 4, a flash evaporator 5, a heating electronic expansion valve 6, a cooling electronic expansion valve 7, a gas replenishment solenoid valve 8, a liquid storage tank 18, and a gas-liquid separator 19.

[0090] Compressor 1 is a low-temperature EVI jet enthalpy-increasing scroll compressor with three interfaces: an exhaust port, an intake port, and a makeup gas port. The exhaust port outputs high-temperature, high-pressure refrigerant gas, the intake port draws in low-temperature, low-pressure refrigerant gas, and the makeup gas port receives medium-pressure makeup refrigerant from flash evaporator 5, thus achieving enthalpy-increasing compression.

[0091] The four-way reversing valve 2 is an electromagnetic reversing valve with four ports: D, C, E, and S. It is used to switch the refrigerant flow direction to achieve the conversion between heating and cooling modes. Port D is connected to the discharge port of compressor 1 to receive high-temperature and high-pressure exhaust gas; Port C is connected to the water heat exchanger 4, which outputs exhaust gas in heating mode and receives return flow in cooling mode; Port E is connected to the finned heat exchanger 3, which outputs exhaust gas in cooling mode and receives return flow in heating mode; Port S is connected to the inlet of the gas-liquid separator 19, and the gas finally returns to the suction port of compressor 1.

[0092] The finned heat exchanger 3 is an air-cooled heat exchanger located on the outdoor side. In heating mode, it acts as an evaporator to absorb heat from the ambient air, and in cooling mode, it acts as a condenser to dissipate heat to the ambient air. An ambient temperature sensor 15 is mounted on its upper part.

[0093] Water heat exchanger 4 is a water-cooled plate heat exchanger located on the indoor side. In heating mode, it acts as a condenser, releasing heat to the water side; in cooling mode, it acts as an evaporator, absorbing heat from the water side. Its inlet is connected to inlet water temperature sensor 13, and its outlet is connected to outlet water temperature sensor 14.

[0094] Flash evaporator 5 is a gas-liquid separation type intercooler with three interfaces: inlet, outlet, and gas makeup port. It is used to realize gas-liquid separation of refrigerant and intermediate gas makeup. The inlet receives medium-pressure refrigerant after it has been throttled by heating electronic expansion valve 6 or refrigeration electronic expansion valve 7. After internal gas-liquid separation, the gas phase is output through the gas makeup port, and the liquid phase is further throttled through the outlet.

[0095] The heating electronic expansion valve 6 is located between the inlet of the flash evaporator 5 and the water heat exchanger 4. In heating mode, it acts as the main throttling element, controlling the refrigerant to be throttled from the condensing pressure to the intermediate pressure, while adjusting the heat exchange on the water side.

[0096] The refrigeration electronic expansion valve 7 is located between the outlet of the flash evaporator 5 and the finned heat exchanger 3. In refrigeration mode, it acts as the main throttling element, controlling the refrigerant to be throttled from the intermediate pressure to the evaporation pressure, while simultaneously regulating the heat exchange on the air side.

[0097] The gas injection solenoid valve 8 is a PWM-controlled electronic expansion valve located between the gas injection port of flash evaporator 5 and the gas injection port of compressor 1. It is used to control the gas injection volume, thereby achieving enthalpy adjustment and exhaust temperature control.

[0098] The liquid storage tank 18 is located between the finned heat exchanger 3 and the gas-liquid separator 19 to store liquid refrigerant and stabilize the system circulation flow.

[0099] The gas-liquid separator 19 is located between the liquid storage tank 18 and the suction port of the compressor 1 to separate the liquid refrigerant in the suction gas and prevent liquid slugging in the compressor.

[0100] The connection relationship of the refrigeration cycle circuit is explained below.

[0101] In heating mode, the refrigerant flow is as follows: compressor 1 discharge port → four-way reversing valve 2 D port → four-way reversing valve 2 C port → water heat exchanger 4 → heating electronic expansion valve 6 → flash evaporator 5 inlet → flash evaporator 5 gas phase via make-up solenoid valve 8 → compressor 1 make-up port, flash evaporator 5 liquid phase → refrigeration electronic expansion valve 7 → finned heat exchanger 3 → four-way reversing valve 2 E port → four-way reversing valve 2 S port → gas-liquid separator 19 → compressor 1 suction port.

[0102] In cooling mode, the refrigerant flow is as follows: compressor 1 discharge port → four-way reversing valve 2 D port → four-way reversing valve 2 E port → finned heat exchanger 3 → refrigeration electronic expansion valve 7 → flash evaporator 5 inlet → flash evaporator 5 gas phase via make-up solenoid valve 8 → compressor 1 make-up port, flash evaporator 5 liquid phase → heating electronic expansion valve 6 → water circuit heat exchanger 4 → four-way reversing valve 2 C port → four-way reversing valve 2 S port → gas-liquid separator 19 → compressor 1 suction port.

[0103] The sensor group includes an exhaust temperature sensor 10, an exhaust pressure sensor 9, and a lubricating oil temperature sensor 16 installed on the compressor 1; an inlet water temperature sensor 13 and an outlet water temperature sensor 14 installed on the water heat exchanger 4; and an ambient temperature sensor 15 installed on the finned heat exchanger 3.

[0104] The exhaust temperature sensor 10 is located at the exhaust port of compressor 1 and is installed close to the exhaust pipe wall to monitor the compressor exhaust temperature Td in real time.

[0105] The exhaust pressure sensor 9 is located on the exhaust pipe between the exhaust port of compressor 1 and the D port of four-way reversing valve 2, and is used to monitor the exhaust pressure in real time and convert it into exhaust saturation temperature Tpd.

[0106] The lubricating oil temperature sensor 16 is located at the bottom of the oil sump of the compressor 1 and is used to monitor the lubricating oil temperature in real time.

[0107] The inlet water temperature sensor 13 is located at the inlet water pipe of the water heat exchanger 4 and is used to monitor the inlet water temperature Tin in real time during the cooling mode.

[0108] The outlet water temperature sensor 14 is located at the outlet water pipe of the water heat exchanger 4 and is used to monitor the outlet water temperature Tout in real time during the heating mode.

[0109] An ambient temperature sensor 15 is located on the upper air inlet side of the finned heat exchanger 3 to monitor the outdoor ambient temperature Tao in real time.

[0110] All sensor signal lines are connected to the analog input port of control module 17 via shielded cables, and the sampling period is synchronized with T1=3s in Example 1.

[0111] The control module 17 is an industrial controller with integrated AI computing capabilities, including a data acquisition unit, an algorithm computing unit, and an instruction output unit.

[0112] The data acquisition unit receives analog signals from all the above sensors, performs filtering, amplification, and analog-to-digital conversion, converts them into digital signals, and stores them in a real-time database.

[0113] The algorithm operation unit has a built-in adaptive threshold prediction model and LSTM time series prediction model. It executes steps S2 to S4 in Example 1 to calculate the dynamic threshold, predict the future exhaust temperature, and generate graded control commands.

[0114] The instruction output unit outputs control instructions via PWM signals and communication bus: it outputs opening adjustment instructions to the heating electronic expansion valve 6, the cooling electronic expansion valve 7 and the gas injection solenoid valve 8; it outputs frequency adjustment instructions to the compressor 1 inverter; and it outputs mode switching instructions to the four-way reversing valve 2.

[0115] The system workflow is described in detail below.

[0116] After the unit is started, the control module 17 first detects the current status of the four-way reversing valve 2 to identify whether the operating mode is heating or cooling. Then, it starts the sensor group to collect all operating parameters at 3-second intervals, including the current exhaust temperature, exhaust temperature change rate, lubricating oil temperature, compressor frequency, ambient temperature, and inlet and outlet water temperatures.

[0117] Based on the collected parameters, control module 17 calls the adaptive threshold prediction model, selects the corresponding coefficient mapping relationship and sign rule according to the current operating mode, and calculates the dynamic threshold of exhaust temperature. At the same time, it calls the LSTM time series prediction model to predict the exhaust temperature for the next 8 seconds based on historical data from the past 10 cycles.

[0118] The control module 17 compares the measured exhaust temperature, the predicted exhaust temperature, and the dynamic threshold, and determines the current control state based on the graded response mechanism: if the warning level condition is met, the opening of the electronic expansion valve is locked and the compressor frequency is optimized; if the intervention level condition is met, the opening of the electronic expansion valve and the amount of supplementary gas are dynamically adjusted according to the temperature change trend; if the recovery level condition is met, the protection state is exited and the system switches to dual-objective optimization control.

[0119] Through the above structure and control process, the system in this embodiment achieves the synergy of multi-parameter fusion acquisition, AI predictive control and adaptive hierarchical adjustment, ensuring accurate and controllable exhaust temperature, efficient unit operation and extended service life.

Claims

1. A control method for an adaptive temperature-controlled chiller system, characterized in that, Includes the following steps: S1, real-time acquisition of the chiller's operating parameters, including at least the current exhaust temperature, exhaust temperature change rate, lubricating oil temperature, compressor frequency, and ambient temperature; S2, based on the current operating mode, call the adaptive threshold prediction model to obtain the dynamic threshold of exhaust temperature. The dynamic threshold of exhaust temperature is obtained by dynamically adjusting the calculation coefficients based on the compressor frequency, ambient temperature and lubricating oil temperature. S3, based on historical operating parameters, predicts the exhaust temperature at future moments; S4, based on the comparison result of the current exhaust temperature, the predicted exhaust temperature and the exhaust temperature too low threshold, adaptively adjust the opening of the electronic expansion valve and / or the compressor frequency.

2. The adaptive temperature control chiller system control method according to claim 1, characterized in that, In step S2, the threshold for excessively low exhaust temperature is determined by the superposition of multiple dynamic correction terms. The dynamic correction terms include at least: a load base term, which is related to the real-time frequency of the compressor; an oil temperature correction term, which is related to the real-time temperature of the lubricating oil to compensate for the risk of lubricating oil flowability; an inlet / outlet water temperature term, which is related to the outlet or inlet water temperature of the water heat exchanger; an ambient temperature term, which is related to the ambient temperature; and a rate of change correction term, which is related to the trend of exhaust temperature change.

3. The adaptive temperature control chiller system control method according to claim 2, characterized in that, In step S2, the load base item, inlet / outlet water temperature item, and ambient temperature item are each configured with dynamic adjustment coefficients to achieve the following: the load base item increases as the compressor load increases; the ambient temperature item increases as the ambient temperature decreases in heating mode and increases as the ambient temperature increases in cooling mode; the inlet / outlet water temperature item increases as the outlet water temperature increases in heating mode and increases as the inlet water temperature decreases in cooling mode.

4. A control method for an adaptive temperature-controlled chiller system according to claim 2 or 3, characterized in that, In step S2, the oil temperature correction term and the rate of change correction term are dynamically determined according to the real-time operating conditions to determine the correction direction and magnitude; in heating mode, the load base term and the inlet and outlet water temperature term have a positive correction effect on the threshold. The ambient temperature term has a negative correction effect on the threshold. In cooling mode, the load base item and the ambient temperature item have a positive correction effect on the threshold. The inlet and outlet water temperature terms have a negative correction effect on the threshold.

5. The adaptive temperature control chiller system control method according to claim 4, characterized in that, In step S2, the correction magnitude of the oil temperature correction item is negatively correlated with the lubricating oil temperature. When the lubricating oil temperature is lower than the normal operating range, the correction magnitude of the oil temperature correction item increases as the temperature decreases.

6. The adaptive temperature control chiller system control method according to claim 4, characterized in that, In step S2, the rate of change correction term is configured with a trend sensitivity coefficient; When the exhaust temperature is decreasing, the trend sensitivity coefficient increases as the absolute value of the rate of decrease increases; when the exhaust temperature is increasing or stable, the trend sensitivity coefficient maintains the baseline value or decreases.

7. The adaptive temperature control chiller system control method according to claim 1, characterized in that, In step S3, the predicted exhaust temperature for future moments is achieved through a time-series prediction model of a long short-term memory network. The historical operating data includes time-series data of exhaust temperature, ambient temperature, and compressor frequency for the past N acquisition cycles. The future moment is the predicted moment 5-10 seconds later.

8. The adaptive temperature control chiller system control method according to claim 1, characterized in that, Step S4 includes: If the exhaust temperature or predicted exhaust temperature is lower than the warning threshold and remains below the first preset duration, the opening of the electronic expansion valve is locked and can only be reduced, and the compressor frequency is reduced. If the exhaust temperature or predicted exhaust temperature is below the intervention threshold and remains below the second preset duration, the opening of the electronic expansion valve is dynamically adjusted according to the temperature change trend. If both the exhaust temperature and the predicted exhaust temperature are higher than the recovery threshold and remain above the third preset duration, the low exhaust temperature control will be exited, and the system will switch to dual-target optimization control of supplementary air temperature and energy consumption.

9. An adaptive temperature-controlled chiller system, employing the adaptive temperature-controlled chiller system control method according to any one of claims 1-8, characterized in that, include: The four-way reversing valve has port C connected to the water heat exchanger, port E connected to the finned heat exchanger, and ports D and S connected to the compressor exhaust port and intake port, respectively. The flash evaporator has its inlet and outlet connected to the water heat exchanger and finned heat exchanger via heating and cooling electronic expansion valves, respectively, and its air supply port is connected to the compressor air supply port via an air supply solenoid valve. The control module is connected to a sensor group that collects signals from the compressor's operating status and environmental parameters, and outputs control commands to the chiller system.

10. An adaptive temperature control chiller system according to claim 9, characterized in that, The sensor group includes: an exhaust temperature sensor, an exhaust pressure sensor, and a lubricating oil temperature sensor installed on the compressor; an inlet water temperature sensor and an outlet water temperature sensor installed on the water heat exchanger; and an ambient temperature sensor installed on the finned heat exchanger. The control module outputs control commands to the compressor, the heating electronic expansion valve, the cooling electronic expansion valve, and the gas injection solenoid valve.

Images