Based on the electron screening effect and the excitation of deuteron harmonic oscillators in a palladium lattice, possible explanations of cold fusion phenomena and the possibility of nuclear heating are discussed. A narrow window is proposed to reach the ∼10 W/cm3 required nuclear heating for three-body fusion by a hypothetical excitation-screening model. A relatively wide window is feasible to reach a few fusion events per second per cubic centimetre under the non-stationary conditions of deuteron charging and discharging. Cold fusion is not feasible under stationary lattice conditions. To confirm the cold fusion phenomena, a heavy water electrolysis experiment is carried out using biased-pulse electrolytic currents, in order to enhance the detection of cold fusion events during charging and discharging of deuterons. A cross-checking system consisting of a recoil-proton scintillation detector and a 3He thermal neutron detector is used to determine the patterns of neutron emission over time. To determine the energy of the emitted neutrons, the pulse-height spectra of the recoil-proton detector are monitored. For a deuterium charging time of 300 h, neutron yields of 1 to 2 n/s·cm3 are obtained for time intervals of 60 to 200 h. From the recoil-proton spectra, it is confirmed that 2.45-MeV neutrons from the D(d, n)3He fusion branch reaction are emitted. The observed time patterns of neutron emission suggest the existence of cold fusion under deuterium charging and discharging.