Residential demand response programs aim to activate demand flexibility at the household level.In recent years,reinforcement learning(RL)has gained significant attention for these type of applications.A major challeng...Residential demand response programs aim to activate demand flexibility at the household level.In recent years,reinforcement learning(RL)has gained significant attention for these type of applications.A major challenge of RL algorithms is data efficiency.New RL algorithms,such as proximal policy optimisation(PPO),have tried to increase data efficiency.Addi tionally,combining RL with transfer learning has been proposed in an effort to mitigate this challenge.In this work,we further improve upon state-of-the-art transfer learning performance by incorporating demand response domain knowledge into the learning pipeline.We evaluate our approach on a demand response use case where peak shaving and self-consumption is incentivised by means of a capacity tariff.We show our adapted version of PPO,combined with transfer learming,reduces cost by 14.51%compared to a regular hysteresis controller and by 6.68%compared to traditional PPO.展开更多
Using Geant4 Monte Carlo code and Technology Computer-Aided Design(TCAD) simulation,energy deposition and charge collection of single event effects(SEE) are studied,which are induced by low-energy protons and α parti...Using Geant4 Monte Carlo code and Technology Computer-Aided Design(TCAD) simulation,energy deposition and charge collection of single event effects(SEE) are studied,which are induced by low-energy protons and α particles in small feature size devices.We analyzed charge collection of SEE especially at Bragg's peak and obtained two types of deposited energy distributions of protons and α particles at different incident energies.The two components of the total charge collected are quantified,which are due to drift current of the space charge region and current in the funnel region separately.Results explain the high soft error rate in experiments of low energy proton.展开更多
Considering the basic heat transfer rules of superfluid helium and the phase diagram of helium, pressure effects on heat transfer to HeⅡ are especially studied. If the bath pressure is less than λ pressure ...Considering the basic heat transfer rules of superfluid helium and the phase diagram of helium, pressure effects on heat transfer to HeⅡ are especially studied. If the bath pressure is less than λ pressure ( P λ ), special peak heat flux density relations are shown to correlate bath pressure, hydrostatic head and modified pressure items (Van der waals pressure and fountain pressure). If the bath pressure is greater than λ pressure( P>P λ), a generalized formula of peak flux density of HeⅡp bath is shown.展开更多
文摘Residential demand response programs aim to activate demand flexibility at the household level.In recent years,reinforcement learning(RL)has gained significant attention for these type of applications.A major challenge of RL algorithms is data efficiency.New RL algorithms,such as proximal policy optimisation(PPO),have tried to increase data efficiency.Addi tionally,combining RL with transfer learning has been proposed in an effort to mitigate this challenge.In this work,we further improve upon state-of-the-art transfer learning performance by incorporating demand response domain knowledge into the learning pipeline.We evaluate our approach on a demand response use case where peak shaving and self-consumption is incentivised by means of a capacity tariff.We show our adapted version of PPO,combined with transfer learming,reduces cost by 14.51%compared to a regular hysteresis controller and by 6.68%compared to traditional PPO.
基金supported by the State Key Program of National Natural Science Foundation of China (Grant No 60836004)the National Natural Science Foundation of China (Grant Nos 61076025 and 61006070)the Specialized Research Fund for the Doctoral Program of Higher Education of China (Grant No 20104307120006)
文摘Using Geant4 Monte Carlo code and Technology Computer-Aided Design(TCAD) simulation,energy deposition and charge collection of single event effects(SEE) are studied,which are induced by low-energy protons and α particles in small feature size devices.We analyzed charge collection of SEE especially at Bragg's peak and obtained two types of deposited energy distributions of protons and α particles at different incident energies.The two components of the total charge collected are quantified,which are due to drift current of the space charge region and current in the funnel region separately.Results explain the high soft error rate in experiments of low energy proton.
基金the National Natural Science Foundation of China
文摘Considering the basic heat transfer rules of superfluid helium and the phase diagram of helium, pressure effects on heat transfer to HeⅡ are especially studied. If the bath pressure is less than λ pressure ( P λ ), special peak heat flux density relations are shown to correlate bath pressure, hydrostatic head and modified pressure items (Van der waals pressure and fountain pressure). If the bath pressure is greater than λ pressure( P>P λ), a generalized formula of peak flux density of HeⅡp bath is shown.
